sas/ets 15.2 user’s guide

SAS/ETS® 15.2User’s GuideIntroduction

SAS® DocumentationAugust 17, 2020

This document is an individual chapter from SAS/ETS® 15.2 User’s Guide.

The correct bibliographic citation for this manual is as follows: SAS Institute Inc. 2020. SAS/ETS® 15.2 User’s Guide. Cary, NC:SAS Institute Inc.

SAS/ETS® 15.2 User’s Guide

Copyright © 2020, SAS Institute Inc., Cary, NC, USA

All Rights Reserved. Produced in the United States of America.

For a hard-copy book: No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or byany means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, SAS InstituteInc.

For a web download or e-book: Your use of this publication shall be governed by the terms established by the vendor at the timeyou acquire this publication.

The scanning, uploading, and distribution of this book via the Internet or any other means without the permission of the publisher isillegal and punishable by law. Please purchase only authorized electronic editions and do not participate in or encourage electronicpiracy of copyrighted materials. Your support of others’ rights is appreciated.

U.S. Government License Rights; Restricted Rights: The Software and its documentation is commercial computer softwaredeveloped at private expense and is provided with RESTRICTED RIGHTS to the United States Government. Use, duplication, ordisclosure of the Software by the United States Government is subject to the license terms of this Agreement pursuant to, asapplicable, FAR 12.212, DFAR 227.7202-1(a), DFAR 227.7202-3(a), and DFAR 227.7202-4, and, to the extent required under U.S.federal law, the minimum restricted rights as set out in FAR 52.227-19 (DEC 2007). If FAR 52.227-19 is applicable, this provisionserves as notice under clause (c) thereof and no other notice is required to be affixed to the Software or documentation. TheGovernment’s rights in Software and documentation shall be only those set forth in this Agreement.

SAS Institute Inc., SAS Campus Drive, Cary, NC 27513-2414

August 2020

SAS® and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in theUSA and other countries. ® indicates USA registration.

Other brand and product names are trademarks of their respective companies.

SAS software may be provided with certain third-party software, including but not limited to open-source software, which islicensed under its applicable third-party software license agreement. For license information about third-party software distributedwith SAS software, refer to http://support.sas.com/thirdpartylicenses.

http://support.sas.com/thirdpartylicenses

Chapter 2

Introduction

ContentsOverview of SAS/ETS Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Uses of SAS/ETS Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Contents of SAS/ETS Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

SAS/ETS High-Performance Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Experimental Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12About This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Chapter Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Syntax Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14Typographical Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Options Used in Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Where to Turn for More Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Accessing the SAS/ETS Sample Library . . . . . . . . . . . . . . . . . . . . . . . . 16SAS Short Courses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16SAS Technical Support Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Major Features of SAS/ETS Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16ARIMA (Box-Jenkins) and ARIMAX (Box-Tiao) Modeling and Forecasting . . . . . 16Structural Time Series Modeling and Forecasting . . . . . . . . . . . . . . . . . . . . 18Regression with Autocorrelated and Heteroscedastic Errors . . . . . . . . . . . . . . 18Count Data Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Multinomial Discrete Choice Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 21Panel Data Linear Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Qualitative and Limited Dependent Variable Analysis . . . . . . . . . . . . . . . . . 23Spatial Econometric Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Vector Time Series Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Simultaneous Systems Linear Regression . . . . . . . . . . . . . . . . . . . . . . . . 26Linear Systems Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Polynomial Distributed Lag Regression . . . . . . . . . . . . . . . . . . . . . . . . . 28Nonlinear Systems Regression and Simulation . . . . . . . . . . . . . . . . . . . . . 29State Space Modeling and Forecasting . . . . . . . . . . . . . . . . . . . . . . . . . 30Spectral Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Distribution of the Severity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Compound Distribution Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Similarity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Seasonal Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34Automatic Time Series Forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Time Series Interpolation and Frequency Conversion . . . . . . . . . . . . . . . . . . 36

6 F Chapter 2: Introduction

Trend and Seasonal Analysis on Transaction Databases . . . . . . . . . . . . . . . . 38Endogeneity and Instrumental Variables . . . . . . . . . . . . . . . . . . . . . . . . . 38Access to Financial and Economic Databases . . . . . . . . . . . . . . . . . . . . . . 40Access to World Weather and NOAA Severe Weather Inventory Databases . . . . . . 45Spreadsheet Calculations and Financial Report Generation . . . . . . . . . . . . . . . 46Loan Analysis, Comparison, and Amortization . . . . . . . . . . . . . . . . . . . . . 47Time Series Forecasting System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48ODS Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Related SAS Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Base SAS Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50SAS Forecast Studio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53SAS/STAT Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53SAS/IML Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54SAS/OR Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55SAS/QC Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55MLE for User-Defined Likelihood Functions . . . . . . . . . . . . . . . . . . . . . . 56JMP Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56SAS Enterprise Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57SAS Add-In for Microsoft Office . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58SAS Enterprise Miner—Time Series Node . . . . . . . . . . . . . . . . . . . . . . . 59

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Overview of SAS/ETS SoftwareSAS/ETS software, a component of the SAS System, provides SAS procedures for the following:

� econometric analysis

� time series analysis

� time series forecasting

� panel data analysis, including dynamic panels

� spatial econometric linear models

� systems modeling and simulation

� discrete choice analysis

� analysis of qualitative and limited dependent variable models

� seasonal adjustment of time series data

� financial analysis and reporting

Uses of SAS/ETS Software F 7

� access to economic and financial databases

� access to global weather databases

� time series data management

� high-performance econometric analysis in symmetric multiprocessing (SMP) mode

In addition to SAS procedures, SAS/ETS software also includes seamless access to economic, financial, andweather databases and interactive environments for time series forecasting and investment analysis.

Uses of SAS/ETS SoftwareSAS/ETS software provides tools for a wide variety of applications in business, government, and academia.Major uses of SAS/ETS procedures are economic analysis, forecasting, economic and financial modeling,time series analysis, financial reporting, and manipulation of time series data.

The common theme relating the many applications of the software is time series data: SAS/ETS software isuseful whenever it is necessary to analyze or predict processes that take place over time or to analyze modelsthat involve simultaneous relationships.

Although SAS/ETS software is most closely associated with business, finance, and economics, time seriesdata also arise in many other fields. SAS/ETS software is useful whenever time dependencies, simultaneousrelationships, or dynamic processes complicate data analysis. For example, an environmental qualitystudy might use SAS/ETS software’s time series analysis tools to analyze pollution emissions data. Apharmacokinetic study might use SAS/ETS software’s features for nonlinear systems to model the dynamicsof drug metabolism in different tissues.

The diversity of problems for which econometrics and time series analysis tools are needed is reflected in theapplications reported by SAS users. The following listed items are some applications of SAS/ETS softwarepresented by SAS users at past annual conferences of the SAS Users Groups (SUGI and SAS Global Forum):

� analyzing heart rate variability of a sleep apnea and cardiovascular patient (Wongdhamma 2016)

� seasonality and interdependence of parking meter transactions (Milhøj 2015)

� modeling operational risk in banking (Rozo, Crook, and Moreira 2015)

� estimating volatility of financial assets (LaBarr 2014)

� analyzing levels, seasonality, and trends in e-commerce (Milhøj 2012)

� early detection of epidemic outbreaks (Shtatland and Shtatland 2008)

� modeling long-run water quality trends (Ragavan and Fernandez 2006)

� neural networks and genetic algorithms for forecasting automobile demand (McNelis and Nickelsburg2002)

� forecasting college enrollment (Calise and Earley 1997)

� fitting a pharmacokinetic model (Morelock et al. 1995)


� testing interaction effects in reducing sudden infant death syndrome (Fleming, Gibson, and Fleming1996)

� forecasting operational indices to measure productivity changes (McCarty 1994)

� spectral decomposition and reconstruction of nuclear plant signals (Hoyer and Gross 1993)

� estimating parameters for the constant-elasticity-of-substitution translog model (Hisnanick 1993)

� applying econometric analysis for mass appraisal of real property (Amal and Weselowski 1993)

� forecasting telephone usage data (Fischetti, Heathcote, and Perry 1993)

� forecasting demand and utilization of inpatient hospital services (Hisnanick 1992)

� using conditional demand estimation to determine electricity demand (Keshani and Taylor 1992)

� estimating tree biomass for measurement of forestry yields (Parresol and Thomas 1991)

� evaluating the theory of input separability in the production function of U.S. manufacturing (Hisnanick1991)

� forecasting dairy milk yields and composition (Benseman 1990)

� predicting the gloss of coated aluminum products subject to weathering (Khan 1990)

� learning curve analysis for predicting manufacturing costs of aircraft (LeBouton 1989)

� analyzing Dow Jones stock index trends (Earley, Sweeney, and Zekavat 1989)

� analyzing the usefulness of the composite index of leading economic indicators for forecasting theeconomy (Lin and Myers 1988)

Contents of SAS/ETS Software

Procedures

SAS/ETS software includes the following SAS procedures:

ARIMA ARIMA (Box-Jenkins) and ARIMAX (Box-Tiao) modeling and forecasting

AUTOREG regression analysis with autocorrelated or heteroscedastic errors and ARCH and GARCHmodeling

COMPUTAB spreadsheet calculations and financial report generation

COPULA fitting and simulating multivariate distributions by using copula methods

COUNTREG regression modeling for dependent variables that represent counts

DATASOURCE access to financial and economic databases

ENTROPY maximum entropy-based regression

ESM forecasting by using exponential smoothing models with optimized smoothing weights

Contents of SAS/ETS Software F 9

EXPAND time series interpolation, frequency conversion, and transformation of time series

LOAN loan analysis and comparison

MDC multinomial discrete choice analysis

MODEL nonlinear simultaneous equations regression and nonlinear systems modeling and simula-tion

PANEL panel data modeling

PDLREG polynomial distributed lag regression

QLIM qualitative and limited dependent variable analysis

SEVERITY modeling the statistical distribution of the severity of losses and other events

SIMILARITY similarity analysis of time series data for time series data mining

SIMLIN linear systems simulation

SPATIALREG spatial econometric models for cross-sectional data

SPECTRA spectral and cross-spectral analysis

SSM state space modeling of time series

STATESPACE state space modeling and automated forecasting of multivariate time series

SYSLIN linear simultaneous equations models

TIMEDATA analyzes time-stamped transactional data with respect to time and accumulates the datainto a time series format

TIMEID identifying the time frequency for data sets that contain time series data

TIMESERIES analysis of time-stamped transactional data

TSCSREG time series cross-sectional regression analysis

UCM unobserved components analysis of time series

VARMAX vector autoregressive and moving average with modeling and forecasting

X11 seasonal adjustment (Census X-11 and X-11 ARIMA)

X12 seasonal adjustment (Census X-12 ARIMA)

X13 seasonal adjustment (Census X-13 ARIMA-SEATS)

High-Performance (HP) Procedures

High-performance (HP) procedures are adapted to perform optimally in symmetric multiprocessing (SMP)mode, providing faster performance by making multiple CPUs available to complete individual processessimultaneously.

SAS/ETS software includes the following high-performance procedures:

HPCDM high-performance compound distribution models

HPCOPULA high-performance fitting and simulation of multivariate distributions by using copulamethods

HPCOUNTREG high-performance regression modeling for count dependent variables


HPPANEL high-performance panel data modeling

HPQLIM high-performance qualitative and limited dependent variable analysis

HPSEVERITY high-performance modeling of the severity of losses and other events

Access Interfaces to Economic and Financial Databases

SAS/ETS software includes the following LIBNAME statement engines to provide access to financial andeconomic databases:

SASECRSP LIBNAME engine for accessing time series and event data that reside in a CRSPAccessdatabase

SASEFAME LIBNAME engine for accessing time series or case series data that reside in a FAMEdatabase

SASEFRED LIBNAME engine to retrieve economic data from the FRED website, which is hosted bythe Economic Research Division of the Federal Reserve Bank of St. Louis

SASEHAVR LIBNAME engine for accessing time series that reside in a Haver Analytics Data LinkExpress (DLX) database

SASEOECD LIBNAME engine for accessing time series to retrieve statistical data from the Organi-sation for Economic Cooperation and Development (OECD) website on topics such asagriculture and fisheries, economy, education, employment, energy, environment, finance,health, industry and entrepreneurship, innovation, insurance and pensions, internationalmigration, internet economy, investment, OECD.Stat data warehouse, regional, rural andurban development, science and technology, social and welfare issues, tax, trade, andtransport

SASEQUAN LIBNAME engine to retrieve economic data from the Quandl website, which offers accessto 8 million time series data sets from 400 sources in finance, economics, society, health,energy, demography, and more

SASEXCCM LIBNAME engine for accessing data items that reside in the CRSP US Stock (STK)Database, the CRSP US Stock and Indices (IND) Database, the CRSP US Treasury (TRS)Database, or the CRSP/Compustat Merged (CCM) Database, which is created from datadelivered via Standard & Poor’s Compustat Xpressfeed product

SASEXFSD LIBNAME engine for accessing both FactSet data and FactSet-sourced data that areprovided by the FactSet OnDemand service

SASEWBGO LIBNAME engine for accessing time series to retrieve statistical data from the WorldBank Group Open (WBGO) data website, hosted by the World Bank Group. The mostpopular is the World Development Indicators (WDI) database, which presents the mostcurrent and accurate global development data available, including national, regional, andglobal estimates. The SASEWBGO interface engine supports access to the WDI database,but it also provides access to time series in other WBGO databases, such as the GlobalEconomic Monitor (GEM) and the Special Data Dissemination Standard (SDDS)

Contents of SAS/ETS Software F 11

Access Interfaces to Global Weather and NOAA Severe Weather Data Inventory Databases

SAS/ETS software includes the following LIBNAME statement engines to provide access to global weatherand severe weather databases:

SASENOAA LIBNAME engine to retrieve severe weather data such as tornado vortex signatures; meso-cyclone signatures; digital mesocyclone detection algorithm; hail, storm cell structure,and preliminary local storm reports; and severe thunderstorm, tornado, flash flood, andspecial marine warnings from the NOAA Severe Weather Data Inventory (SWDI) webservice

SASERAIN LIBNAME engine to retrieve global weather data such as temperature, precipitation(rainfall), weather description, weather icon, and wind speed from the World WeatherOnline website

Macros

SAS/ETS software includes the following SAS macros:

%AR generates statements to define autoregressive error models for the MODEL procedure

%EQAR defines autoregressive error models that are specified using general form equations for theMODEL procedure

%BOXCOXAR investigates Box-Cox transformations useful for modeling and forecasting a time series

%DFPVALUE computes probabilities for Dickey-Fuller test statistics

%DFTEST performs Dickey-Fuller tests for unit roots in a time series process

%LOGTEST tests to determine whether a log transformation is appropriate for modeling and forecastinga time series

%MA generates statements to define moving-average error models for the MODEL procedure

%EQMA defines moving-average error models that are specified using general form equations forthe MODEL procedure

%PDL generates statements to define polynomial distributed lag models for the MODEL proce-dure

These macros are part of the SAS AUTOCALL facility and are automatically available for use in your SASprogram. For information about the SAS macro facility, see SAS Macro Language: Reference.

The Time Series Forecasting System

SAS/ETS software includes an interactive forecasting system, described in Part IV. This graphical userinterface to SAS/ETS forecasting features was developed with SAS/AF software and uses PROC ARIMA andother internal routines to perform time series forecasting. The Time Series Forecasting System makes it easyto forecast time series and provides many features for graphical data exploration and graphical comparisonsof forecasting models and forecasts. (You must have SAS/GRAPH installed to use the graphical features ofthe system.)

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=mcrolref&docsetTarget=titlepage.htm


SAS/ETS High-Performance ProceduresSAS/ETS high-performance procedures provide econometric modeling tools that have been specially devel-oped to take advantage of parallel processing in both multithreaded single-machine mode and distributedmultiple-machine mode. You can run all these procedures in single-machine mode without licensing SASHigh-Performance Econometrics. However, to run these procedures in distributed mode, you must licenseSAS High-Performance Econometrics.

Econometric modeling methods available in high-performance environment include regression for countdata, models for the severity of losses or other events, compound distribution modeling, regression modelsfor qualitative and limited dependent variables, copula simulation, and panel data modeling. In addition tothe high-performance econometric procedures described in this book, SAS/ETS includes high-performanceutility procedures, which are described in Base SAS Procedures Guide: High-Performance Procedures.

Experimental SoftwareExperimental software is sometimes included as part of a production-release product. It is provided tocustomers in order to obtain feedback. All experimental features are marked Experimental in this document.Whenever an experimental procedure, statement, or option is used, a message is displayed in the SAS log toindicate that it is experimental. The design and syntax of experimental software might change before anyproduction release. Experimental software has been tested prior to release, but it has not necessarily beentested to production-quality standards, so it should be used with care.

About This BookThis book is a user’s guide to SAS/ETS software. Since SAS/ETS software is a part of the SAS System, thisbook assumes that you are familiar with Base SAS software and have the books SAS Programmers Guide:Essentials and Base SAS Procedures Guide available for reference. It also assumes that you are familiar withSAS data sets, the SAS DATA step, and with basic SAS procedures such as PROC PRINT and PROC SORT.Chapter 4, “Working with Time Series Data,” in this book summarizes the aspects of Base SAS software thatare most relevant to the use of SAS/ETS software.

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=prochp&docsetTarget=titlepage.htm

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=lepg&docsetTarget=titlepage.htm


https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=proc&docsetTarget=titlepage.htm

Chapter Organization F 13

Chapter OrganizationFollowing a brief What’s New, this book is divided into five major parts. Part I contains general informationto aid you in working with SAS/ETS Software. Part II explains the SAS procedures of SAS/ETS software.Part III describes the available data access interfaces for economic, financial and weather databases. Finally,Part IV is the reference for the Time Series Forecasting System, an interactive forecasting menu system thatuses PROC ARIMA and other routines to perform time series forecasting.

The new features added to SAS/ETS software since the publication of SAS/ETS Software: Changes andEnhancements for Release 13.2 are summarized in Chapter 1, “What’s New in SAS/ETS 15.2.” If you haveused SAS/ETS software in the past, you may want to skim this chapter to see what’s new.

Part I contains the following chapters.

Chapter 2, the current chapter, provides an overview of SAS/ETS software and summarizes related SASpublications, products, and services.

Chapter 4, “Working with Time Series Data,” discusses the use of SAS data management and programmingfeatures for time series data.

Chapter 5, “Date Intervals, Formats, and Functions,” summarizes the time intervals, date and datetimeinformats, date and datetime formats, and date and datetime functions available in the SAS System.

Chapter 6, “SAS Macros and Functions,” documents SAS macros and DATA step financial functions providedwith SAS/ETS software. The macros use SAS/ETS procedures to perform Dickey-Fuller tests, test for theneed for log transformations, or select optimal Box-Cox transformation parameters for time series data.

Chapter 7, “Nonlinear Optimization Methods,” documents the NonLinear Optimization subsystem used bysome ETS procedures to perform nonlinear optimization tasks.

Part II contains chapters that explain the SAS procedures that make up SAS/ETS software. These chaptersappear in alphabetical order by procedure name.

Part III contains chapters that document the ETS access interfaces to economic, financial and weatherdatabases.

Each of the chapters that document the SAS/ETS procedures (Part II) and the SAS/ETS access interfaces(Part III) is organized as follows:

1. The “Overview” section gives a brief description of the procedure.

2. The “Getting Started” section provides a tutorial introduction on how to use the procedure.

3. The “Syntax” section is a reference to the SAS statements and options that control the procedure.

4. The “Details” section discusses various technical details.

5. The “Examples” section contains examples of the use of the procedure.

6. The “References” section contains technical references on methodology.

Part IV contains the chapters that document the features of the Time Series Forecasting System.


Syntax ConventionsEach procedure’s “Syntax” section follows the conventions that are described in this section. Consider thefollowing statements:

CLASS variable < (options) > . . . < variable < (options) > > < / global-options > ;

RANGE FROM from TO to ;

<label:> TEST <'string'> equation1 < , equation2. . . > / test-options ;

These statements demonstrate the syntax conventions that are described in the following list:

UPPERCASE BOLD is used for keywords in lists of SAS statements and options in “Syntax” sections.When you type a keyword in SAS code, you type it as shown (although any mixof uppercase and lowercase is valid). In the preceding examples, the statementnames (CLASS, RANGE, and TEST) are keywords. In addition, the FROM andTO are required keywords in the RANGE statement. Note that keywords aredisplayed only in uppercase (not bold) when they are used in text.

oblique is used in syntax definitions and in text to represent arguments for which yousupply a value. The preceding CLASS statement indicates that variable, options,and global-options are arguments for which you can supply values. The valuesthat you can supply are defined later in the description of the CLASS statement.

< > (angle brackets) identify optional arguments. Arguments that are not enclosed inangle brackets are required. In the preceding CLASS statement, you must supplya value for one variable because the first variable is not enclosed in angle brackets.However, supplying values for additional variables, options, and global-optionsis optional.

. . . (ellipsis dots) indicate that the preceding argument can be repeated. In thepreceding CLASS statement, the “. . . ” indicates that you can supply additionalvariables, (along with optional options). Sometimes the argument is shown againafter the “. . . ” to emphasize that it can be repeated.

'value' (straight quotes around a value) indicate that the value must be enclosed inquotation marks (which can be single or double quotes). In the preceding TESTstatement, straight quotes around string indicate that you must use quotationmarks when you specify a string.

( ) (parentheses) indicate arguments that must be grouped together. In the precedingCLASS statement, you must type parentheses around the options in order toindicate which syntax elements are options and which are variables. Statementsthat do not require parentheses to indicate association sometimes allow you toomit the parentheses when you specify only one option; these cases are indicatedin the statement description.

| (vertical bar) indicates that you can choose one value from a group of values.Values that are separated by a vertical bar are mutually exclusive. A vertical barindicates mutually exclusive values for an option or indicates aliases for an optionname.

; (semicolon) indicates the end of a statement.

Typographical Conventions F 15

Other special characters—such as an equal sign (=), tilde (�), colon (:), and slash (/)—indicate where in thesyntax you must type those characters.

Typographical ConventionsThis book uses several type styles for presenting information. The following list explains the meaning of thetypographical conventions used in this book:

UPPERCASE ROMAN is used for SAS statements, options, and other SAS language elements whenthey appear in the text. However, you can enter these elements in your own SASprograms in lowercase, uppercase, or a mixture of the two.

VariableName is used for the names of variables and data sets when they appear in the text.

bold is used to refer to matrices and vectors.

italic is used for terms that are defined in the text, for emphasis, and for references topublications.

monospace is used for example code. In most cases, this book uses lowercase type for SAScode.

Options Used in ExamplesThe HTMLBLUE style is used to create the graphs and the HTML tables that appear in the online documen-tation. The PEARLJ style is used to create the PDF tables that appear in the documentation. A style templatecontrols stylistic elements such as colors, fonts, and presentation attributes. You can specify a style templatein an ODS destination statement as follows:

ods html style=HTMLBlue;. . .ods html close;

ods pdf style=PearlJ;. . .ods pdf close;

Most of the PDF tables are produced by using the following SAS System option:

options papersize=(6.5in 9in);

If you run the examples, you might get slightly different output. This is a function of the SAS System optionsthat are used and the precision that your computer uses for floating-point calculations.


Where to Turn for More InformationThis section describes other sources of information about SAS/ETS software.

Accessing the SAS/ETS Sample LibraryThe SAS/ETS Sample Library includes many examples that illustrate the use of SAS/ETS software, includingthe examples used in this documentation. To access these sample programs, select Help from the menuand then select SAS Help and Documentation. From the Contents list, select the section Sample SASPrograms under Learning to Use SAS.

SAS Short CoursesThe SAS Education Division offers a number of training courses that might be of interest to SAS/ETS users.Please check the SAS web site for the current list of available training courses.

SAS Technical Support ServicesAs with all SAS products, the SAS Technical Support staff is available to respond to problems and answertechnical questions regarding the use of SAS/ETS software.

Major Features of SAS/ETS SoftwareThe following sections summarize major features of SAS/ETS software. For more information, see thechapters on individual procedures.

ARIMA (Box-Jenkins) and ARIMAX (Box-Tiao) Modeling and ForecastingThe ARIMA procedure provides the identification, parameter estimation, and forecasting of autoregressiveintegrated moving-average (Box-Jenkins) models, seasonal ARIMA models, transfer function models, andintervention models. The ARIMA procedure includes the following features:

� complete ARIMA (Box-Jenkins) modeling with no limits on the order of autoregressive or moving-average processes

� model identification diagnostics, including the following:

– autocorrelation function

– partial autocorrelation function

ARIMA (Box-Jenkins) and ARIMAX (Box-Tiao) Modeling and Forecasting F 17

– inverse autocorrelation function

– cross-correlation function

– extended sample autocorrelation function

– minimum information criterion for model identification

– squared canonical correlations

� stationarity tests

� outlier detection

� intervention analysis

� regression with ARMA errors

� transfer function modeling with fully general rational transfer functions

� seasonal ARIMA models

� ARIMA model-based interpolation of missing values

� several parameter estimation methods, including the following:

– exact maximum likelihood

– conditional least squares

– exact nonlinear unconditional least squares (ELS or ULS)

� prewhitening transformations

� forecasts and confidence limits for all models

� forecasting tied to parameter estimation methods: finite memory forecasts for models estimated bymaximum likelihood or exact nonlinear least squares methods and infinite memory forecasts for modelsestimated by conditional least squares

� diagnostic statistics to help judge the adequacy of the model, including the following:

– Akaike’s information criterion (AIC)

– Schwarz’s Bayesian criterion (SBC or BIC)

– Box-Ljung chi-square test statistics for white-noise residuals

– autocorrelation function of residuals

– partial autocorrelation function of residuals

– inverse autocorrelation function of residuals

– automatic outlier detection


Structural Time Series Modeling and ForecastingThe UCM procedure provides a flexible environment for analyzing time series data using structural time seriesmodels, also called unobserved components models (UCM). These models represent the observed series as asum of suitably chosen components such as trend, seasonal, cyclical, and regression effects. You can use theUCM procedure to formulate comprehensive models that bring out all the salient features of the series underconsideration. Structural models are applicable in the same situations where Box-Jenkins ARIMA modelsare applicable; however, the structural models tend to be more informative about the underlying stochasticstructure of the series. The UCM procedure includes the following features:

� general unobserved components modeling where the models can include trend, multiple seasons andcycles, and regression effects

� maximum-likelihood estimation of the model parameters

� model diagnostics that include a variety of goodness-of-fit statistics, and extensive graphical diagnosisof the model residuals

� forecasts and confidence limits for the series and all the model components

� Model-based seasonal decomposition

� extensive plotting capability that includes the following:

– forecast and confidence interval plots for the series and model components such as trend, cycles,and seasons

– diagnostic plots such as residual plot, residual autocorrelation plots, and so on

– seasonal decomposition plots such as trend, trend plus cycles, trend plus cycles plus seasons, andso on

� model-based interpolation of series missing values

� full sample (also called smoothed) estimates of the model components

Regression with Autocorrelated and Heteroscedastic ErrorsThe AUTOREG procedure provides regression analysis and forecasting of linear models with autocorrelatedor heteroscedastic errors. The AUTOREG procedure includes the following features:

� estimation and prediction of linear regression models with autoregressive errors

� autoregressive or subset autoregressive processes of any order

� optional stepwise selection of autoregressive parameters

� choice of the following estimation methods:

– exact maximum likelihood

Regression with Autocorrelated and Heteroscedastic Errors F 19

– exact nonlinear least squares

– Yule-Walker

– iterated Yule-Walker

� tests for any linear hypothesis that involves the structural coefficients

� restrictions for any linear combination of the structural coefficients

� forecasts with confidence limits

� estimation and forecasting for A of ARCH (autoregressive conditional heteroscedasticity), and thefollowing variations:

– GARCH (generalized autoregressive conditional heteroscedasticity)

– IGARCH (integrated GARCH)

– EGARCH (exponential GARCH)

– QGARCH (quadratic GARCH)

– TGARCH (threshold GARCH)

– PGARCH (power GARCH)

– GARCH-M (GARCH-in-mean)

� combination of ARCH and GARCH models with autoregressive models, with or without regressors

� estimation and testing of general heteroscedasticity models

� variety of model diagnostic information, including the following:

– autocorrelation plots

– partial autocorrelation plots

– Durbin-Watson test statistic and generalized Durbin-Watson tests of any order

– Durbin h and Durbin t statistics

– Godfrey LM test

– Ramsey’s RESET test

– McLeod-Li portmanteau Q test for ARCH disturbances

– Engle’s LM test for ARCH disturbances

– Lee and King’s for ARCH disturbances

– Wong and Li’s test for ARCH disturbances

– Chow test

– Bai-Perron supF, UDmaxF, WDmaxF, and supF(l C 1jl) tests

– Akaike’s information criterion

– Schwarz information criterion

– Phillips-Perron stationarity test

– Phillips-Ouliaris cointegration test


– Kwiatkowski-Phillips-Schmidt-Shin (KPSS) test

– Shin cointegration test

– augmented Dickey-Fuller test

– Engle-Granger cointegration test

– Elliot, Rothenberg, and Stock test

– Ng and Perron test

– tests for statistical independence

– Jarque-Bera test for normality

– CUSUM and CUMSUMSQ statistics

� exact significance levels (p-values) for the Durbin-Watson statistic

� embedded missing values

Count Data ModelsThe COUNTREG procedure provides regression models in which the dependent variable takes nonnegativeinteger count values. The COUNTREG procedure supports the following features:

� Poisson regression

� Conway-Maxwell-Poisson regression

� negative binomial regression with quadratic and linear variance functions

� zero-inflated Poisson (ZIP) regression

� zero-inflated Conway-Maxwell-Poisson regression

� zero-inflated negative binomial (ZINB) regression

� fixed- and random-effects Poisson panel data models

� fixed- and random-effects NB (negative binomial) panel data models

� variable selection

� Bayesian estimation and inference, including diagnostic plots

Multinomial Discrete Choice Analysis F 21

Multinomial Discrete Choice AnalysisThe MDC procedure provides maximum likelihood (ML) or simulated maximum likelihood estimates ofmultinomial discrete choice models in which the choice set consists of unordered multiple alternatives. Thedecision makers can be people, households, firms, or any other decision-making units, and the alternativesare a set of competing options. Unordered multiple choices are observed in many settings, including choicesof housing location, occupation, political party affiliation, and mode of transportation.

The MDC procedure supports the following models and features:

� intuitive

� conditional logit

� nested logit

� heteroscedastic extreme value

� multinomial probit

� mixed logit

� pseudorandom or quasi-random numbers for simulated maximum likelihood estimation

� bounds imposed on the parameter estimates

� linear restrictions imposed on the parameter estimates

� SAS data set containing predicted probabilities and linear predictor (x0ˇ) values

� decision tree and nested logit

� model fit and goodness-of-fit measures, including the following:

– likelihood ratio

– Aldrich-Nelson

– Cragg-Uhler 1

– Cragg-Uhler 2

– Estrella

– adjusted Estrella

– McFadden’s LRI

– Veall-Zimmermann


– Schwarz criterion or Bayesian information criterion (BIC)


Panel Data Linear ModelsThe PANEL procedure deals with panel data sets that consist of time series observations on each of severalcross-sectional units. The PANEL procedure includes the following features:

� one-way and two-way fixed effects

� one-way and two-way random effects

� variance component estimation by the following methods:

– Fuller and Battese method (variance component model)

– Wansbeek and Kapteyn method

– Wallace and Hussain method

– Nerlove method

� Parks method (autoregressive model)

� Da Silva method (mixed variance component moving-average model)

� Hausman-Taylor and Amemiya-MaCurdy estimation

� dynamic-panel estimation one-step, two-step, or iterative generalized method of moments (GMM)

� support for unbalanced panel data for all methods

� model specification tests

� panel data unit-root tests

� automatic generation of lagged variables

� model comparison tables

� model specification tests

� variety of estimates and statistics, including the following:

– underlying error components estimates

– regression parameter estimates

– standard errors of estimates

– t tests

– R-square statistic

– correlation matrix of estimates

– covariance matrix of estimates

– autoregressive parameter estimate

– cross-sectional components estimates

– autocovariance estimates

– F tests of linear hypotheses about the regression parameters

– specification tests, including the Hausman test

Qualitative and Limited Dependent Variable Analysis F 23

Qualitative and Limited Dependent Variable AnalysisThe QLIM procedure analyzes univariate and multivariate limited dependent variable models where dependentvariables take discrete values or dependent variables are observed only in a limited range of values. Thisprocedure includes logit, probit, Tobit, and general simultaneous equations models. The QLIM procedureincludes the following features:

� linear regression with heteroscedasticity

� probit models with heteroscedasticity

� logit models with heteroscedasticity

� Tobit models (censored and truncated) with heteroscedasticity

� Box-Cox regression with heteroscedasticity

� bivariate probit models

� bivariate Tobit models

� ordered logit and ordered probit models

� sample selection models, including the Heckman model

� multivariate limited dependent models

� stochastic frontier models

� random effects and random coefficients

� Bayesian estimation and inference, including diagnostic plots

� residual plots, predictive plots, marginal-effects plots, and so on

Spatial Econometric ModelsThe SPATIALREG procedure analyzes spatial econometric models for cross-sectional data where observationsare spatially referenced or georeferenced. The SPATIALREG procedure includes the following features:

� linear models with spatial log of X (SLX) effects

� spatial autoregressive (SAR) model

� spatial Durbin model (SDM)

� spatial error model (SEM)

� spatial Durbin error model (SDEM)


� spatial moving average (SMA) model

� spatial Durbin moving average (SDMA) model

� spatial autoregressive moving average (SARMA) model

� spatial Durbin autoregressive moving average (SDARMA) model

� spatial autoregressive confused (SAC) model

� spatial Durbin autoregressive confused (SDAC) model

� k-order binary contiguity spatial weight matrices

� k-order nearest neighbor spatial weight matrices

� compact representations of spatial weight matrices

� Taylor and Chebyshev approximations for large data sets

Vector Time Series AnalysisThe VARMAX procedure enables you to model the dynamic relationship both between the dependentvariables and between the dependent and independent variables. The VARMAX procedure includes thefollowing features:

� several modeling features:

– vector autoregressive model (VAR)

– vector autoregressive model with exogenous variables (VARX)

– vector autoregressive and moving-average model (VARMA)

– vector autoregressive and moving-average model with exogenous variables (VARMAX)

– vector autoregressive fractionally integrated moving-average model (VARFIMA)

– vector autoregressive fractionally integrated moving-average model with exogenous variables(VARFIMAX)

– Bayesian vector autoregressive model (BVAR)

– vector error correction model (VECM)

– Bayesian vector error correction model (BVECM)

– GARCH-type multivariate conditional heteroscedasticity models (BEKK, CCC, DCC)

– vector error correction model in ARMA-GARCH form

� criteria for automatically determining AR and MA orders:


– corrected AIC (AICC)

– Hannan-Quinn (HQ) criterion

Vector Time Series Analysis F 25

– final prediction error (FPE)

– Schwarz Bayesian criterion (SBC), also known as Bayesian information criterion (BIC)

� AR order identification aids:

– partial cross-correlations

– Yule-Walker estimates

– partial autoregressive coefficients

– partial canonical correlations

� testing the presence of unit roots and cointegration:

– Dickey-Fuller tests

– Johansen cointegration test for nonstationary vector processes of integrated order one

– Stock-Watson common trends test for the possibility of cointegration among nonstationary vectorprocesses of integrated order one

– Johansen cointegration test for nonstationary vector processes of integrated order two

� model parameter estimation methods:

– least squares (LS)

– maximum likelihood (ML)

– conditional maximum likelihood (CML)

� model checks and residual analysis using the following tests:

– Durbin-Watson (DW) statistics

– F test for autoregressive conditional heteroscedastic (ARCH) disturbance

– F test for AR disturbances

– Jarque-Bera normality test

– portmanteau test

� seasonal deterministic terms

� subset models

� multiple regression with distributed lags

� dead-start model that does not have present values of the exogenous variables

� Granger-causal relationships between two distinct groups of variables

� infinite order AR representation

� impulse response function (or infinite order MA representation)

� decomposition of the predicted error covariances

� roots of the characteristic functions for both the AR and MA parts to evaluate the proximity of theroots to the unit circle


� contemporaneous relationships among the components of the vector time series

� forecasts

� conditional covariances for GARCH models

� log-likelihood output

� specification of initial parameter values for optimization

� constraints and bounds on parameters for optimization

� Wald tests

Simultaneous Systems Linear RegressionThe SYSLIN and ENTROPY procedures provide regression analysis of a simultaneous system of linearequations.

The SYSLIN procedure includes the following features:

� estimation of parameters in simultaneous systems of linear equations

� full range of estimation methods including the following:

– ordinary least squares (OLS)

– two-stage least squares (2SLS)

– three-stage least squares (3SLS)

– iterated 3SLS (IT3SLS)

– seemingly unrelated regression (SUR)

– iterated SUR (ITSUR)

– limited-information maximum likelihood (LIML)

– full-information maximum likelihood (FIML)

– minimum expected loss (MELO)

– general K-class estimators

� weighted regression

� any number of restrictions for any linear combination of coefficients, within a single model or acrossequations

� tests for any linear hypothesis, for the parameters of a single model or across equations

� wide range of model diagnostics and statistics including the following:

– usual ANOVA tables and R-square statistics

– Durbin-Watson statistics

Simultaneous Systems Linear Regression F 27

– standardized coefficients

– test for overidentifying restrictions

– residual plots

– standard errors and t tests

– covariance and correlation matrices of parameter estimates and equation errors

� predicted values, residuals, parameter estimates, and variance-covariance matrices saved in output SASdata sets

� other features of the SYSLIN procedure that enable you to do the following:

– impose linear restrictions on the parameter estimates

– test linear hypotheses about the parameters

– write predicted and residual values to an output SAS data set

– write parameter estimates to an output SAS data set

– write the crossproducts matrix (SSCP) to an output SAS data set

– use raw data, correlations, covariances, or cross products as input

The ENTROPY procedure supports the following models and features:

� generalized maximum entropy (GME) estimation

� generalized cross entropy (GCE) estimation

� normed moment generalized maximum entropy

� maximum entropy-based seemingly unrelated regression (MESUR) estimation

� pure inverse estimation

� estimation of parameters in simultaneous systems of linear equations

� Markov models

� unordered multinomial choice problems

� weighted regression

� any number of restrictions for any linear combination of coefficients, within a single model or acrossequations

� tests for any linear hypothesis, for the parameters of a single model or across equations


Linear Systems SimulationThe SIMLIN procedure performs simulation and multiplier analysis for simultaneous systems of linearregression models. The SIMLIN procedure includes the following features:

� reduced form coefficients

� interim multipliers

� total multipliers

� dynamic multipliers

� multipliers for higher-order lags

� dynamic forecasts and simulations

� goodness-of-fit statistics

� acceptance of the equation system coefficients estimated by the SYSLIN procedure as input

Polynomial Distributed Lag RegressionThe PDLREG procedure provides regression analysis for linear models with polynomial distributed (Almon)lags. The PDLREG procedure includes the following features:

� entry of any number of regressors as a polynomial lag distribution and the use of any number ofcovariates

� use of any order lag length and degree polynomial for lag distribution

� optional upper and lower endpoint restrictions

� specification of any number of linear restrictions on covariates

� option to repeat analysis over a range of degrees for the lag distribution polynomials

� support for autoregressive errors to any lag

� forecasts with confidence limits

Nonlinear Systems Regression and Simulation F 29

Nonlinear Systems Regression and SimulationThe MODEL procedure provides parameter estimation, simulation, and forecasting of dynamic nonlinearsimultaneous equation models. The MODEL procedure includes the following features:

� nonlinear regression analysis for systems of simultaneous equations, including weighted nonlinearregression

� full range of parameter estimation methods including the following:

– nonlinear ordinary least squares (OLS)

– nonlinear seemingly unrelated regression (SUR)

– nonlinear two-stage least squares (2SLS)

– nonlinear three-stage least squares (3SLS)

– iterated SUR

– iterated 3SLS

– generalized method of moments (GMM)

– nonlinear full-information maximum likelihood (FIML)

– simulated method of moments (SMM)

� supports dynamic multi-equation nonlinear models of any size or complexity

� uses the full power of the SAS programming language for model definition, including left-hand-sideexpressions

� hypothesis tests of nonlinear functions of the parameter estimates

� linear and nonlinear restrictions of the parameter estimates

� bounds imposed on the parameter estimates

� computation of estimates and standard errors of nonlinear functions of the parameter estimates

� estimation and simulation of ordinary differential equations (ODEs), and differential algebraic equa-tions (DAEs)

� vector autoregressive error processes and polynomial lag distributions easily specified for the nonlinearequations

� variance modeling (ARCH, GARCH, and others)

� computation of goal-seeking solutions of nonlinear systems to find input values needed to producetarget outputs

� dynamic, static, or n-period-ahead forecast simulation modes

� simultaneous solution or single equation solution modes

� Monte Carlo simulation using parameter estimate covariance and across-equation residuals covariancematrices or user-specified random functions


� Monte Carlo simulation of multidimensional systems using copulas

� a variety of diagnostic statistics including the following

– model R-square statistics– general Durbin-Watson statistics and exact p-values– asymptotic standard errors and t tests– first-stage R-square statistics– covariance estimates– collinearity diagnostics– simulation goodness-of-fit statistics– Theil inequality coefficient decompositions– Theil relative change forecast error measures– heteroscedasticity tests– Godfrey test for serial correlation– Hausman specification test– Chow tests

� block structure and dependency structure analysis for the nonlinear system

� listing and cross-reference of fitted model

� automatic calculation of needed derivatives by using exact analytic formula

� efficient sparse matrix methods used for model solution; choice of other solution methods

Model definition, parameter estimation, simulation, and forecasting can be performed interactively in a singleSAS session, or models can be stored in files and reused and combined in later runs.

State Space Modeling and ForecastingThe SSM procedure provides state space modeling of univariate and multivariate time series and longitudinaldata. State space models encompass an alternative general formulation of multivariate ARIMA models. TheSSM procedure includes the following features:

� general linear state space models (SMMs)

� expressive language to specify an SSM, including flexible and intuitive specification of transition andcovariance matrices

� easy specification of commonly used SSMs by using only a few keywords

� restricted maximum likelihood estimation computed using the (diffuse) Kalman filter algorithm

� forecasts, residuals, and full-sample estimations of any linear combination of state variables

� residual diagnostics plots

� plots for detecting structural breaks

Spectral Analysis F 31

Spectral AnalysisThe SPECTRA procedure provides spectral analysis and cross-spectral analysis of time series. The SPECTRAprocedure includes the following features:

� efficient calculation of periodogram and smoothed periodogram using fast finite Fourier transform andChirp-Z algorithms

� multiple spectral analysis, including raw and smoothed spectral and cross-spectral function estimates,with user-specified window weights

� choice of kernel for smoothing

� output of the following spectral estimates to a SAS data set:

– Fourier sine and cosine coefficients

– periodogram

– smoothed periodogram

– cospectrum

– quadrature spectrum

– amplitude

– phase spectrum

– squared coherency

� Fisher’s Kappa and Bartlett’s Kolmogorov-Smirnov test statistic for testing a null hypothesis of whitenoise

Distribution of the SeverityThe SEVERITY procedure estimates parameters of any probability distribution that is used to model themagnitude (severity) of a continuous-valued event of interest. The SEVERITY procedure includes thefollowing features:

� parameter estimation of predefined distribution models, including the following:

– Burr distribution

– exponential distribution

– gamma distribution

– generalized Pareto distribution

– inverse Gaussian (Wald) distribution

– lognormal distribution

– Pareto distribution


– Tweedie distribution

– Weibull distribution

� parameter estimation of arbitrarily defined parametric distribution models

� fitting distributions to data by either truncation or censoring

� group estimation

� several fit statistics, including the following:

– log likelihood


– corrected Akaike’s information criterion (AICC)

– Schwarz Bayesian information criterion (BIC)

– Kolmogorov-Smirnov statistic (KS)

– Anderson-Darling statistic (AD)

– Cramér–von Mises statistic (CvM)

� regression effects

� scoring functions

� multithreaded computation

� ability to specify the objective function for optimization

� plots of the estimated cumulative distribution function (CDF), the estimated empirical distributionfunction (EDF), and the estimated probability density function (PDF)

Compound Distribution ModelsThe HPCDM procedure computes an estimate of the compound distribution model, given the distributions ofthe parameters. For example, PROC HPCDM can estimate the distribution of the aggregate loss during atime period of interest, given the distribution models of the frequency (count) and of the severity of loss.

The HPCDM procedure includes the following features:

� accepts severity models estimated by the SEVERITY procedure and frequency models estimated bythe COUNTREG procedure

� scenario analysis with regression effects

� group scenario analysis with classification and interaction effects

� support for externally simulated counts

� parameter perturbation analysis that assesses the effect of parameter uncertainty associated withfrequency and severity models

� ability to compute the distribution of aggregate adjusted loss

Similarity Analysis F 33

Similarity AnalysisThe SIMILARITY procedure computes similarity measures associated with time-stamped data, time series,and other sequentially ordered numeric data. The SIMILARITY procedure includes the following features:

� ability to accumulate time-stamped data into a time series

� missing value interpretation

� zero value interpretation

� functional transformations of time series, including the following:

– log (LOG)

– square-root (SQRT)

– logistic (LOGISTIC)

– Box-Cox (BOXCOX)

– user-defined transformations

� simple differencing and seasonal differencing

� time series missing value trimming

� time warping by compressing or expanding the input sequence with respect to the target sequence

� sequence normalizations, including the following:

– standard (STANDARD)

– absolute (ABSOLUTE)

– user-defined normalizations

� sequence scaling, including the following:

– standard (STANDARD)

– absolute (ABSOLUTE)

– user-defined scaling

� ability to compute similarity measures, including the following:

– squared deviation (SQRDEV)

– absolute deviation (ABSDEV)

– mean square deviation (MSQRDEV)

– mean absolute deviation (MABSDEV)

– user-defined similarity measures

� sliding similarity measures analysis with three types of sequence sliding:

– no sliding


– slide by time index

– slide by season index

� support for large data sets

Seasonal AdjustmentThe X13 procedure provides seasonal adjustment of time series by using the US Bureau of the CensusX-13ARIMA-SEATS seasonal adjustment program. The X-13ARIMA-SEATS program was developed bythe Time Series Staff of the Statistical Research Division, US Census Bureau, by incorporating the SEATSmethod into the X-12-ARIMA seasonal adjustment program.

The X13 procedure generalizes the older X11 and X12 procedures and includes the following features:

� US Bureau of the Census X-13ARIMA-SEATS seasonal adjustment program

� support for the X-12 ARIMA method

� support for the X-11 ARIMA method

� all the features of the Census Bureau program

� processing of any number of variables at once with no maximum length for a series

� decomposition of monthly or quarterly series into seasonal, trend, trading day, and irregular components

� multiplicative, additive, pseudo-additive, and log additive forms of the decomposition

� support for regARIMA modeling

� automatic identification of outliers

� support for TRAMO-based automatic model selection

� support for sliding spans analysis

� use of regressors to process missing values within the span of the series

� computation of tests for stable, moving, and combined seasonality

� spectral analysis of original, seasonally adjusted, and irregular series

� ability to project seasonal component one year ahead, which enables reintroduction of seasonal factorsfor an extrapolated series

� full control over what is printed or output

Automatic Time Series Forecasting F 35

Automatic Time Series ForecastingThe ESM procedure provides a quick way to generate forecasts for many time series or transactional data inone step by using exponential smoothing methods. All parameters associated with the forecasting model areoptimized based on the data.

You can use the following smoothing models:

� simple

� double

� linear

� damped trend

� seasonal

� Winters method (additive and multiplicative)

Additionally, PROC ESM can transform the data before applying the smoothing methods using any of thesetransformations:

� log

� square root

� logistic

� Box-Cox

In addition to forecasting, the ESM procedure can also produce graphic output.

The ESM procedure can forecast both time series data, whose observations are equally spaced at a specifictime interval (for example, monthly, weekly), or transactional data, whose observations are not spaced withrespect to any particular time interval. (Internet, inventory, sales, and similar data are typical examples oftransactional data. For transactional data, the data are accumulated based on a specified time interval to forma time series.)


Time Series Interpolation and Frequency ConversionThe EXPAND procedure provides time interval conversion and missing value interpolation for time series.The EXPAND procedure includes the following features:

� conversion of time series frequency; for example, constructing quarterly estimates from annual seriesor aggregating quarterly values to annual values

� conversion of irregular observations to periodic observations

� interpolation of missing values in time series

� conversion of observation types; for example, estimate stocks from flows and vice versa. All possibleconversions are supported between any of the following:

– beginning of period

– end of period

– period midpoint

– period total

– period average

� conversion of time series phase shift; for example, conversion between fiscal years and calendar years

� identifying observations including the following:

– identification of the time interval of the input values

– validation of the input data set observations

– computation of the ID values for the observations in the output data set

� choice of four interpolation methods:

– cubic splines

– linear splines

– step functions

– simple aggregation

� ability to perform extrapolation by a linear projection of the trend of the cubic spline curve fit to theinput data

� ability to transform series before and after interpolation (or without interpolation) by using any of thefollowing:

– constant shift or scale

– sign change or absolute value

– logarithm, exponential, square root, square, logistic, inverse logistic

– lags, leads, differences

– classical decomposition

Time Series Interpolation and Frequency Conversion F 37

– bounds, trims, reverse series

– centered moving, cumulative, or backward moving average

– centered moving, cumulative, or backward moving range

– centered moving, cumulative, or backward moving geometric mean

– centered moving, cumulative, or backward moving maximum

– centered moving, cumulative, or backward moving median

– centered moving, cumulative, or backward moving minimum

– centered moving, cumulative, or backward moving product

– centered moving, cumulative, or backward moving corrected sum of squares

– centered moving, cumulative, or backward moving uncorrected sum of squares

– centered moving, cumulative, or backward moving rank

– centered moving, cumulative, or backward moving standard deviation

– centered moving, cumulative, or backward moving sum

– centered moving, cumulative, or backward moving median

– centered moving, cumulative, or backward moving t-value

– centered moving, cumulative, or backward moving variance

� support for a wide range of time series frequencies:

– YEAR

– SEMIYEAR

– QUARTER

– MONTH

– SEMIMONTH

– TENDAY

– WEEK

– WEEKDAY

– DAY

– HOUR

– MINUTE

– SECOND

� support for repeating of shifting the basic interval types to define a great variety of different frequencies,such as fiscal years, biennial periods, work shifts, and so forth

For more information about time series data transformations, see Chapter 4, “Working with Time SeriesData,” and Chapter 5, “Date Intervals, Formats, and Functions.”


Trend and Seasonal Analysis on Transaction DatabasesThe TIMESERIES procedure can accumulate transactional data to time series and perform trend and seasonalanalysis on the accumulated time series.

Time series analyses performed by the TIMESERIES procedure include the follows:

� descriptive statistics relevant for time series data

� seasonal decomposition and seasonal adjustment analysis

� correlation analysis

� cross-correlation analysis

The TIMESERIES procedure includes the following features:

� ability to process large amounts of time-stamped transactional data

� statistical methods useful for large-scale time series analysis or (temporal) data mining

� output data sets stored in either a time series format (default) or a coordinate format (transposed)

The TIMESERIES procedure is normally used to prepare data for subsequent analysis that uses otherSAS/ETS procedures or other parts of the SAS system. The time series format is most useful when the dataare to be analyzed with SAS/ETS procedures. The coordinate format is most useful when the data are tobe analyzed with SAS/STAT procedures or SAS Enterprise Miner. (For example, clustering time-stampedtransactional data can be achieved by using the results of TIMESERIES procedure with the clusteringprocedures of SAS/STAT and the nodes of SAS Enterprise Miner.)

Endogeneity and Instrumental VariablesSAS/ETS software provides several procedures that estimate models that have endogeneity. Endogeneityusually occurs for three reasons: omitted variables, measurement error in regressors, and simultaneity. Indynamic models, endogeneity is even more relevant, because regressors might be correlated with the errorterm not only from the current time period but from preceding periods as well. The following proceduressupport models that have endogeneity.

The MODEL procedure includes the following features related to endogeneity:

� nonlinear regression analysis of single equations

� nonlinear regression analysis of systems of simultaneous equations

� support for general-form models that have endogeneity

� a variety of estimation methods to handle endogeneity, including the following:

– (nonlinear) two-stage least squares (2SLS)

Endogeneity and Instrumental Variables F 39

– iterated two-stage least squares (IT2SLS)

– (nonlinear) three-stage least squares (3SLS)

– iterated three-stage least squares (IT3SLS)

– generalized method of moments (GMM)

– iterated generalized method of moments (ITGMM)


The SYSLIN procedure includes the following features related to endogeneity:

� linear regression analysis of single equations

� linear regression analysis of systems of simultaneous equations

� a variety of estimation methods to handle endogeneity, including the following:

– (nonlinear) two-stage least squares (2SLS)

– (nonlinear) three-stage least squares (3SLS)

– iterated three-stage least squares (IT3SLS)

– limited-information maximum likelihood (LIML)

– minimum expected loss (MELO)

– general K-class estimators


The SIMLIN procedure performs simulation and multiplier analysis of simultaneous systems of linearregression models that have endogeneity.

The QLIM procedure includes the following features related to endogeneity:

� test of endogeneity for a list of regressors in the model

� overidentification test for the validity of instrumental variables

� ability to estimate models that have endogeneity by adding regressions of endogenous regressors onexogenous regressors and instrumental variables

� ability to estimate structural models that contain one endogenous variable by using full-informationmaximum likelihood (FIML)

� ability to estimate structural models that contain multiple endogenous variables by using simulatedmaximum likelihood

The PANEL procedure uses instrumental variable regressions to estimate both static and dynamic panelmodels that have endogeneity:

� Hausman-Taylor and Amemiya-MaCurdy estimation for static panel models

� One-step, two-step, or iterative generalized method of moments (GMM) for dynamic panel models


Access to Financial and Economic DatabasesThe DATASOURCE procedure and the SAS/ETS data access interface LIBNAME engines (SASECRSP,SASEFAME, SASEFRED, SASEHAVR, SASEOECD, SASEQUAN, SASEWBGO, SASEXCCM andSASEXFSD) provide seamless, efficient access to time series data from data files supplied by a variety ofcommercial and governmental data vendors.

The DATASOURCE procedure includes the following features:

� support for data files distributed by the following data vendors:

– DRI/McGraw-Hill

– FAME Information Services

– Haver Analytics

– Standard & Poor’s Compustat Service

– Center for Research in Security Prices (CRSP)

– International Monetary Fund

– US Bureau of Labor Statistics

– US Bureau of Economic Analysis

– Organization for Economic Cooperation and Development (OECD)

� ability to select the series, frequency, time range, and cross sections of extracted data

� ability to create an output data set containing descriptive information about the series available in thedata file

� ability to read EBCDIC data on ASCII systems and vice versa

The SASECRSP interface LIBNAME engine includes the following features:

� enables random access to time series data residing in CRSPAccess databases

� provides a seamless interface between CRSP and SAS data processing

� uses the LIBNAME statement to enable you to specify which time series you want to read from theCRSPAccess database and how you want to perform selection

� enables you access to CRSP Stock, CRSP/COMPUSTAT Merged (CCM), or CRSP Indices Data

� provides convenient formats, informats, and functions for CRSP and SAS datetime conversions

The SASEFAME interface LIBNAME engine includes the following features:

� provides SAS and FAME users with flexibility in accessing and processing time series data, case series,and formulas that reside in either a FAME database or a SAS data set

Access to Financial and Economic Databases F 41

� uses the LIBNAME statement to enable you to specify which time series you want to read from theFAME database

� enables you to convert the selected time series to the same time scale

� works with the SAS DATA step to perform further subsetting and to store the resulting time series in aSAS data set

� performs more analysis if desired in either the same SAS session or a later session

� supports the FAME CROSSLIST function for subsetting via BY groups

� supports the use of FAME in a client/server environment

� enables access to your FAME remote data when you specify the port number of the TCP/IP servicethat is defined for your FAME Master server and the node name of your FAME master server in yourSASEFAME libref’s physical path

The SASEFRED interface LIBNAME engine includes the following features:

� enables SAS users to retrieve economic data from the FRED website, which is hosted by the EconomicResearch Division of the Federal Reserve Bank of St. Louis

� provides access to various sources of FRED data, including those from Dow Jones & Company andthe Federal Reserve System

� provides query options that allow you to request information by date, series, source, release, tag, orcategory

� enables selection of time series variables that you want to read into SAS based on a list of IDs thatname the index or series

� defines the range of observations based on a specified date range or a specified offset and limit (cutoff)

� aggregates the selected time series to a specified aggregation frequency and specified aggregationmethod

� supports TLS connectivity by obtaining a secure connection using the CONNECT method (if necessary)and a PROXY

� creates an XML map of the data for dynamic, flexible association of SAS formats and informats for allvariables

� supports various data transformations, including rates of change

� enables you to select the vintage dates you want to use when accessing archival (ALFRED) time series

The SASEHAVR interface LIBNAME engine includes the following features:

� gives Windows users random access to economic and financial data residing in a Haver Analytics DataLink Express (DLX) database


� provides the following types of Haver data sets:

– US Economic Indicators

– Specialized Databases

– Financial Indicators

– Industry

– Industrial Countries

– Emerging Markets

– International Organizations

– Forecasts and As Reported Data

– United States Regional

� enables you to limit the range of data that is read from the time series

� enables you to specify a desired conversion frequency. Start dates are recommended in the LIBNAMEstatement to help you save resources when processing large databases or when processing a largenumber of observations.

� enables you to use the WHERE, KEEP, or DROP statement in your DATA step to further subset yourdata

� supports use of the SQL procedure to create a view of your resulting SAS data set

The SASEOECD interface LIBNAME engine includes the following features:

� enables SAS users to retrieve time series data from the Organization for Economic Cooperation andDevelopment (OECD) web site which offers access to statistical data on topics such as agricultureand fisheries, economy, education, employment, energy, environment, finance, health, industry andentrepreneurship, innovation, insurance and pensions, international migration, internet economy,investment, OECD.Stat data warehouse, regional, rural and urban development, science and technology,social and welfare issues, tax, trade, and transport

� uses the LIBNAME statement to enable you to specify which time series you want to retrieve based onthe data set id and the key sets that you specify

� enables you to limit the time range of data that is retrieved by specifying a start date and an end date

� reads the JSON data into a SAS data set, and automatically maps the JSON data for dynamic, flexibleassociation of SAS formats and informats for all variables


The SASEQUAN interface LIBNAME engine includes the following features:

� enables SAS users to retrieve economic and other time series data from the Quandl website, whichoffers access to over 8 million time series data sets from 400 sources in finance, economics, society,health, energy, demography, and more

Access to Financial and Economic Databases F 43

� provides various sources of QUANDL data, including those from NASDAQ, Merrill Lynch, NikkeiGroup, the Wall Street Journal, Google Finance, Yahoo Finance, and various foreign and domesticstock and commodity exchanges

� uses the LIBNAME statement to enable you to specify which time series you want to read fromQUANDL

� enables selection of time series variables that you want to read into SAS based on a list of QUANDLcodes that name the index or series

� defines the range of observations based on a specified date range

� sorts the order of observation in either ascending or descending time order

� enables you to collapse the selected time series to the same frequency

� supports various data transformations, including those that accumulate or difference the series




The SASEWBGO interface LIBNAME engine includes the following features:

� enables SAS programmers to retrieve time series data from the World Bank Group Open (WBGO)data website, hosted by the World Bank Group

� uses the LIBNAME statement to enable you to specify how to retrieve your WBGO data

� enables selection of time series data that you want to read into SAS based on a list of country codesthat name the countries whose data you want to read

� enables selection of time series variables that you want to read into SAS based on a list of time seriesindicator codes that name the series

� defines the range of observations based on a range of years, and an optional page number and numberof observations per page to report

� sorts the order of observations in ascending or descending time order

� provides a utility data set, XWBGOTPU, containing useful information (downloaded from a specifiedURL) about countries based on income level, time series indicators based on source ID, or time seriesindicators based on topic ID




The SASEXCCM interface LIBNAME engine includes the following features:

� enables random access to time series data residing in CRSPAccess databases

� provides a seamless interface between CRSP, Compustat XpressFeed, and SAS data processing

� uses the LIBNAME statement to enable you to specify which data items, data groups, and time seriesyou want to read from the CRSPAccess database and how you want to perform selection

� supports data-item-handling access methods to CRSP Stock (STK), CRSP/COMPUSTAT Merged(CCM), CRSP Indices (IND), or CRSP Treasury (TRS) DData

� provides selection based on keys such as GVKEY, PERMNO, INDNO, TREASNO, and TCUSIP forefficient access to data items

The SASEXFSD interface LIBNAME engine includes the following features:

� enables SAS users to access both FactSet data and FactSet-sourced data that are provided by theFactSet OnDemand service (formerly known as FASTFetch)

� uses the LIBNAME statement to specify which factlet (provided by FactSet) to use to open a FactSetdatabase and to select the desired access method for subsetting and selecting data

� provides updated access to various sources of FactSet OnDemand offerings for financial data, includingcommodity benchmarks, banking data, and broker research

� works with the SAS DATA step to write the selected FactSet data to a SAS data set

� enables you to specify a range of dates for time series selection by either relative or absolute dates

� enables you to specify a FactSet frequency for displaying the data by using any of over 20 availablecodes

� provides TLS connectivity by obtaining a secure connection using the CONNECT method (if necessary)and a PROXY

� allows for ECON_EXPR_DATA and FQL (FactSet Query Language) syntax for function returns fromFactSet

� allows for SPEC_ID_DATA and FQL economic download syntax


Access to World Weather and NOAA Severe Weather Inventory Databases F 45

Access to World Weather and NOAA Severe Weather Inventory DatabasesThe SAS/ETS data access interface LIBNAME engines (SASERAIN and SASENOAA) provide seamless,efficient access to weather events and weather time series data supplied by World Weather Online and theNOAA Severe Weather Data Inventory web services.

The SASENOAA interface LIBNAME engine includes the following features:

� enables SAS users to access severe weather data sets, such as those for tornado vortex signatures(NX3TVS), storm cell structure (NX3STRUCTURE), and preliminary local storm reports (PLSR)

� works with the SAS DATA step to write the selected NOAA data to a SAS data set

� selects data based on geospatial limits, such as by a bounding box or a centerpoint-radius combination

� selects data based on a date range

� returns data in these formats:

– XML; data are returned in XML format

– KMZ; data are returned in zipped KML format for Google My Maps (plot data on a map)

– SHP; mapping data are returned in zipped Esri format (four files returned inside ZIP file)




The SASERAIN interface LIBNAME engine includes the following features:

� enables SAS users to retrieve weather data from the World Weather Online website

� uses the LIBNAME statement to enable you to download World Weather Online data and to specifywhich weather data time series you want to retrieve based on up to nine locations

� works with the SAS DATA step to write the selected weather data to a SAS data set

� selects past weather data based on a date range that starts no earlier than July 1, 2008

� selects local forecast data based on a range defined by number of days (starts today), returns up to 15days of premium local weather forecast data.

� enables you to select the frequency of data, whether daily, hourly, every three hours, or otherwise

� maintains the sort order, so the locations (q-codes) are sorted in the resulting SAS data set by the orderspecified in the QUERY= option, by date (time ID), and by variable (time series item name)


� works with the SAS DATA step to perform further subsetting and to store weather data in a SAS dataset



Spreadsheet Calculations and Financial Report GenerationThe COMPUTAB procedure generates tabular reports using a programmable data table.

The COMPUTAB procedure is especially useful when you need both the power of a programmable spread-sheet and a report-generation system and you want to set up a program to run in batch mode and generateroutine reports. The COMPUTAB procedure includes the following features:

� report generation facility for creating tabular reports such as income statements, balance sheets, andother row and column reports for analyzing business or time series data

� ability to tailor report format to almost any desired specification

� use of the SAS programming language to provide complete control of the calculation and format ofeach item of the report

� ability to report definition in terms of a data table on which programming statements operate

� ability for a single reference to a row or column to bring the entire row or column into a calculation

� ability to create new rows and columns (such as totals, subtotals, and ratios) with a single programmingstatement

� access to individual table values when needed

� built-in features to provide consolidation reports over summarization variables

Loan Analysis, Comparison, and Amortization F 47

Loan Analysis, Comparison, and AmortizationThe LOAN procedure provides analysis and comparison of mortgages and other installment loans; it includesthe following features:

� ability to specify contract terms for any number of different loans and ability to analyze and comparevarious financing alternatives

� analysis of four different types of loan contracts including the following:

– fixed rate

– adjustable rate

– buy-down rate

– balloon payment

� full control over adjustment terms for adjustable rate loans: life caps, adjustment frequency, andmaximum and minimum rates

� support for a wide variety of payment and compounding intervals

� ability to incorporate initialization costs, discount points, down payments, and prepayments (uniformor lump-sum) in loan calculations

� analysis of different rate adjustment scenarios for variable rate loans including the following:

– worst case

– best case

– fixed rate case

– estimated case

� ability to make loan comparisons at different points in time

� ability to make loan comparisons at each analysis date on the basis of five different economic criteria:

– present worth of cost (net present value of all payments to date)

– true interest rate (internal rate of return to date)

– current periodic payment

– total interest paid to date

– outstanding balance

� ability to base loan comparisons on either after-tax or before-tax analysis

� report of the best alternative when loans of equal amount are compared

� amortization schedules for each loan contract

� output that shows payment dates, rather than just payment sequence numbers, when starting date isspecified


� optional printing or output of the amortization schedules, loan summaries, and loan comparisoninformation to SAS data sets

� ability to specify rounding of payments to any number of decimal places

Time Series Forecasting SystemSAS/ETS software includes the Time Series Forecasting System, a point-and-click application for exploringand analyzing univariate time series data. You can use the automatic model selection facility to select thebest-fitting model for each time series, or you can use the system’s diagnostic features and time seriesmodeling tools interactively to develop forecasting models customized to best predict your time series. Thesystem provides both graphical and statistical features to help you choose the best forecasting method foreach series.

The system can be invoked by selecting AnalysisISolutions, by the FORECAST command, and by clickingthe Forecasting icon in the Data Analysis folder of the SAS Desktop.

The following is a brief summary of the features of the Time Series Forecasting system. With the system youcan:

� use a wide variety of forecasting methods, including several kinds of exponential smoothing models,Winters method, and ARIMA (Box-Jenkins) models. You can also produce forecasts by combining theforecasts from several models.

� use predictor variables in forecasting models. Forecasting models can include time trend curves,regressors, intervention effects (dummy variables), adjustments you specify, and dynamic regression(transfer function) models.

� view plots of the data, predicted versus actual values, prediction errors, and forecasts with confidencelimits. You can plot changes or transformations of series, zoom in on parts of the graphs, or plotautocorrelations.

� use hold-out samples to select the best forecasting method

� compare goodness-of-fit measures for any two forecasting models side-by-side or list all models sortedby a particular fit statistic

� view the predictions and errors for each model in a spreadsheet or view and compare the forecastsfrom any two models in a spreadsheet

� examine the fitted parameters of each forecasting model and their statistical significance

� control the automatic model selection process: the set of forecasting models considered, the goodness-of-fit measure used to select the best model, and the time period used to fit and evaluate models

� customize the system by adding forecasting models for the automatic model selection process and forpoint-and-click manual selection

� save your work in a project catalog

ODS Graphics F 49

� print an audit trail of the forecasting process

� save and print system output including spreadsheets and graphs

ODS GraphicsMany SAS/ETS procedures produce graphical output using the SAS Output Delivery System (ODS). TheODS Graphics system provides several advantages:

� Plots and graphs are output objects in the Output Delivery System (ODS) and can be manipulated withODS commands.

� There is no need to write SAS/GRAPH statements or use special plotting macros.

� There are multiple formats to choose from: html, gif, and rtf.

� Templates control the appearance of plots.

� Styles control the color scheme.

� You can edit or create templates and styles for all graphs.

To enable graphical output from SAS/ETS procedures, you must use the following statement in your SASprogram.

ods graphics on;

The graphical output produced by many SAS/ETS procedures can be controlled using the PLOTS= option inthe PROC statement.

For more information about the features of the ODS Graphics system, including the many ways that youcan control or customize the plots produced by SAS procedures, see Chapter 23, “Statistical Graphics UsingODS” (SAS/STAT User’s Guide). For more information about the SAS Output Delivery system, see the SASOutput Delivery System: User’s Guide.

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=statug&docsetTarget=titlepage.htm

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=odsug&docsetTarget=titlepage.htm

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=odsug&docsetTarget=titlepage.htm


Related SAS SoftwareMany features not found in SAS/ETS software are available in other parts of the SAS System, such as BaseSAS, SAS Forecast Server, SAS/STAT software, SAS/OR software, SAS/QC software, SAS Stat Studio, andSAS/IML software.

If you do not find something you need in SAS/ETS software, you might be able to find it in SAS/STATsoftware and in Base SAS software. If you still do not find it, look in other SAS software products or contactSAS Technical Support staff.

The following subsections summarize the features of other SAS products that might be of interest to users ofSAS/ETS software.

Base SAS SoftwareThe features provided by SAS/ETS software are extensions to the features provided by Base SAS software.Many data management and reporting capabilities you need are part of Base SAS software. For documentationof Base SAS software, see SAS Programmers Guide: Essentials and Base SAS Procedures Guide. In particular,for information about statistical analysis features included with Base SAS, see Base SAS Procedures Guide:Statistical Procedures.

The following sections summarize Base SAS software features of interest to users of SAS/ETS software.For further discussion of some of these topics as they relate to time series data and SAS/ETS software, seeChapter 4, “Working with Time Series Data.”

SAS DATA Step

The DATA step is your primary tool for reading and processing data in the SAS System. The DATA stepprovides a powerful general purpose programming language that enables you to perform all kinds of dataprocessing tasks. The DATA step is documented in Base SAS Procedures Guide.

Base SAS Procedures

Base SAS software includes many useful SAS procedures, which are documented in Base SAS ProceduresGuide and Base SAS Procedures Guide: Statistical Procedures. The following is a list of Base SAS proceduresyou might find useful:

CATALOG for managing SAS catalogs

CHART for printing charts and histograms

COMPARE for comparing SAS data sets

CONTENTS for displaying the contents of SAS data sets

COPY for copying SAS data sets

CORR for computing correlations

CPORT for moving SAS data libraries between computer systems

DATASETS for deleting or renaming SAS data sets



https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=procstat&docsetTarget=titlepage.htm






Base SAS Software F 51

FCMP for compiling functions for use in SAS programs. The SAS Function Compiler Procedure(FCMP) enables you to create, test, and store SAS functions and subroutines before youuse them in other SAS procedures. PROC FCMP accepts slight variations of DATA stepstatements, and most features of the SAS programing language can be used in functionsand subroutines that are processed by PROC FCMP.

FREQ for computing frequency crosstabulations

MEANS for computing descriptive statistics and summarizing or collapsing data over cross sections

PLOT for printing scatter plots

PRINT for printing SAS data sets

PROTO for accessing external functions from the SAS system. The PROTO procedure enablesyou to register external functions that are written in the C or C++ programming languages.You can use these functions in SAS as well as in C-language structures and types. Afterthe C-language functions are registered in PROC PROTO, they can be called from anySAS function or subroutine that is declared in the FCMP procedure, as well as from anySAS function, subroutine, or method block that is declared in the COMPILE procedure.

RANK for computing rankings or order statistics

SORT for sorting SAS data sets

SQL for processing SAS data sets with Structured Query Language

STANDARD for standardizing variables to a fixed mean and variance

TABULATE for printing descriptive statistics in tabular format

TIMEPLOT for plotting variables over time

TRANSPOSE for transposing SAS data sets

UNIVARIATE for computing descriptive statistics

Global Statements

Global statements can be specified anywhere in your SAS program, and they remain in effect until changed.Global statements are documented in Base SAS Procedures Guide. You may find the following SAS globalstatements useful:

FILENAME for accessing data files

FOOTNOTE for printing footnote lines at the bottom of each page

%INCLUDE for including files of SAS statements

LIBNAME for accessing SAS data libraries

OPTIONS for setting various SAS system options

QUIT for ending an interactive procedure step

RUN for executing the preceding SAS statements

TITLE for printing title lines at the top of each page

X for issuing host operating system commands from within your SAS session



Some Base SAS statements can be used with any SAS procedure, including SAS/ETS procedures. Thesestatements are not global, and they affect only the SAS procedure they are used with. These statements aredocumented in Base SAS Procedures Guide.

The following Base SAS statements are useful with SAS/ETS procedures:

BY for computing separate analyses for groups of observations

FORMAT for assigning formats to variables

LABEL for assigning descriptive labels to variables

WHERE for subsetting data to restrict the range of data processed or to select or exclude observa-tions from the analysis

SAS Functions

SAS functions can be used in DATA step programs and in the COMPUTAB and MODEL procedures. Thefollowing kinds of functions are available:

� character functions for manipulating character strings

� date and time functions for performing date and calendar calculations

� financial functions for performing financial calculations such as depreciation, net present value, periodicsavings, and internal rate of return

� lagging and differencing functions for computing lags and differences

� mathematical functions for computing data transformations and other mathematical calculations

� probability functions for computing quantiles of statistical distributions and the significance of teststatistics

� random number functions for simulation experiments

� sample statistics functions for computing means, standard deviations, kurtosis, and so forth

SAS functions are documented in Base SAS Procedures Guide. Chapter 4, “Working with Time Series Data,”discusses the use of date, time, lagging, and differencing functions. Chapter 5, “Date Intervals, Formats, andFunctions,” contains a reference list of date and time functions.

Formats, Informats, and Time Intervals

Base SAS software provides formats to control the printing of data values, informats to read data values, andtime intervals to define the frequency of time series. For more information, see Chapter 5, “Date Intervals,Formats, and Functions.”



SAS Forecast Studio F 53

SAS Forecast StudioSAS Forecast Studio is part of the SAS Forecast Server product. It provides an interactive environment formodeling and forecasting very large collections of hierarchically organized time series, such as SKUs inproduct lines and sales regions of a retail business. SAS Forecast Studio greatly extends the capabilitiesprovided by the Time Series Forecasting System included with SAS/ETS and described in Part IV.

SAS Forecast Studio is documented in SAS Forecast Studio: User’s Guide.

SAS/STAT SoftwareSAS/STAT software is of interest to users of SAS/ETS software because many econometric and otherstatistical methods not included in SAS/ETS software are provided in SAS/STAT software.

SAS/STAT software includes procedures for a wide range of statistical methodologies including the following:

� logistic regression

� censored regression

� principal component analysis

� structural equation models using covariance structure analysis

� factor analysis

� survival analysis

� discriminant analysis

� cluster analysis

� categorical data analysis; log-linear and conditional logistic models

� general linear models

� mixed linear and nonlinear models

� generalized linear models

� response surface analysis

� kernel density estimation

� LOESS regression

� spline regression

� two-dimensional kriging

� multiple imputation for missing values

� survey data analysis


SAS/IML SoftwareSAS/IML software gives you access to a powerful and flexible programming language (Interactive MatrixLanguage) in a dynamic, interactive environment. The fundamental object of the language is a data matrix.You can use SAS/IML software interactively (at the statement level) to see results immediately, or youcan store statements in a module and execute them later. The programming is dynamic because necessaryactivities such as memory allocation and dimensioning of matrices are done automatically.

You can access built-in operators and call routines to perform complex tasks such as matrix inversion oreigenvector generation. You can define your own functions and subroutines using SAS/IML modules. Youcan perform operations on an entire data matrix. You have access to a wide choice of data managementcommands. You can read, create, and update SAS data sets from inside SAS/IML software without everusing the DATA step.

SAS/IML software is of interest to users of SAS/ETS software because it enables you to program your owneconometric and time series methods in the SAS System. It contains subroutines for time series operatorsand for general function optimization. If you need to perform a statistical calculation not provided as anautomated feature by SAS/ETS or other SAS software, you can use SAS/IML software to program the matrixequations for the calculation.

Kalman Filtering and Time Series Analysis in SAS/IML

SAS/IML software includes CALL routines and functions for Kalman filtering and time series analysis,which perform the following:

� generate univariate, multivariate, and fractional time series

� compute likelihood function of ARMA, VARMA, and ARFIMA models

� compute an autocovariance function of ARMA, VARMA, and ARFIMA models

� check the stationarity of ARMA and VARMA models

� filter and smooth time series models using Kalman method

� fit AR, periodic AR, time-varying coefficient AR, VAR, and ARFIMA models

� handle Bayesian seasonal adjustment models

SAS/OR Software F 55

SAS/OR SoftwareSAS/OR software provides SAS procedures for operations research and project planning and includes a menudriven system for project management. SAS/OR software has features for the following:

� solving transportation problems

� linear, integer, and mixed-integer programming

� nonlinear programming and optimization

� scheduling projects

� plotting Gantt charts

� drawing network diagrams

� solving optimal assignment problems

� network flow programming

SAS/OR software might be of interest to users of SAS/ETS software for its mathematical programmingfeatures. In particular, the NLP and OPTMODEL procedures in SAS/OR software solve nonlinear program-ming problems and can be used for constrained and unconstrained maximization of user-defined likelihoodfunctions.

For more information, see SAS/OR User’s Guide: Mathematical Programming.

SAS/QC SoftwareSAS/QC software provides a variety of procedures for statistical quality control and quality improvement.SAS/QC software includes procedures for the following:

� Shewhart control charts

� cumulative sum control charts

� moving average control charts

� process capability analysis

� Ishikawa diagrams

� Pareto charts

� experimental design

SAS/QC software also includes the SQC menu system for interactive application of statistical quality controlmethods and the ADX Interface for experimental design.

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=v_001&docsetId=ormpug&docsetTarget=titlepage.htm


MLE for User-Defined Likelihood FunctionsThere are several SAS procedures that enable you to do maximum likelihood estimation of parameters in anarbitrary model with a likelihood function that you define: PROC MODEL, PROC NLP, PROC OPTMODELand PROC IML.

The MODEL procedure in SAS/ETS software enables you to minimize general log-likelihood functions forthe error term of a model.

The NLP and OPTMODEL procedures in SAS/OR software are general nonlinear programming proceduresthat can maximize a general function subject to linear equality or inequality constraints. You can use PROCNLP or OPTMODEL to maximize a user-defined nonlinear likelihood function.

You can use the IML procedure in SAS/IML software for maximum likelihood problems. The optimizationroutines used by PROC NLP are available through IML subroutines. You can write the likelihood function inthe SAS/IML matrix language and call the constrained and unconstrained nonlinear programming subroutinesto maximize the likelihood function with respect to the parameter vector.

JMP SoftwareJMP software uses a flexible graphical interface to display and analyze data. JMP dynamically links statisticsand graphics so you can easily explore data, make discoveries, and gain the knowledge you need to makebetter decisions. JMP provides a comprehensive set of statistical tools as well as design of experiments(DOE) and advanced quality control (QC and SPC) tools for Six Sigma in a single package. JMP is softwarefor interactive statistical graphics and includes the following:

� a data table window for editing, entering, and manipulating data

� a broad range of graphical and statistical methods for data analysis

� a facility for grouping data and computing summary statistics

� JMP scripting language (JSL)—a scripting language for saving and creating frequently used routines

� JMP automation

� Formula Editor—a formula editor for each table column to compute values as needed

� linear models, correlations, and multivariate

� design of experiments module

� options to highlight and display subsets of data

� statistical quality control and variability charts—special plots, charts, and communication capabilityfor quality-improvement techniques

� survival analysis

� time series analysis, which includes the following:

SAS Enterprise Guide F 57

– Box-Jenkins ARIMA forecasting

– seasonal ARIMA forecasting

– transfer function modeling

– smoothing models: Winters method, single, double, linear, damped trend linear, and seasonalexponential smoothing

– diagnostic charts (autocorrelation, partial autocorrelation, and variogram) and statistics of fit

– a model comparison table to compare all forecasts generated

– spectral density plots and white noise tests

� tools for printing and for moving analyses results between applications

SAS Enterprise GuideSAS Enterprise Guide has the following features:

� integration with the SAS9 platform:

– open metadata repository (OMR) integration

– SAS report integration

� create report interface� ODS support� Web report studio integration

– access to information maps

– ETL studio impact analysis

– ESRI integration within the OLAP analyzer

– data mining scoring task

� the user interface and workflow

– process flow

– ability to create stored processes from process flows

– SAS folders window

– project parameters

– query builder interface

– code node

– OLAP analyzer

� ESRI integration� tree-diagram-based OLAP explorer� SAS report snapshots� SAS Web OLAP viewer for .NET ability to create EG projects


– workspace maximization

With SAS Enterprise Guide, you can perform time series analysis with the following procedures:

� prepare time series data—the Prepare Time Series Data task can be used to make data more suitablefor analysis by other time series tasks.

� create time series data—the Create Time Series Data wizard helps you convert transactional data intofixed-interval time series. Transactional data are time-stamped data collected over time with irregularor varied frequency.

� ARIMA Modeling and Forecasting task

� Basic Forecasting task

� Regression Analysis with Autoregressive Errors

� Regression Analysis of Panel Data

SAS Add-In for Microsoft OfficeThe main time series tasks in SAS Add-In for Microsoft Office (AMO) are as follows:

� Prepare Time Series Data

� Basic Forecasting

� ARIMA Modeling and Forecasting

� Regression Analysis with Autoregressive Errors

� Regression Analysis of Panel Data

� Create Time Series Data

� Forecast Studio Create Project

� Forecast Studio Open Project

� Forecast Studio Submit Overrides

SAS Enterprise Miner—Time Series node F 59

SAS Enterprise Miner—Time Series NodeSAS Enterprise Miner is the SAS solution for data mining, streamlining the data mining process to createhighly accurate predictive and descriptive models. SAS Enterprise Miner’s process flow diagram eliminatesthe need for manual coding and reduces the model development time for both business analysts and statisti-cians. The system is customizable and extensible; users can integrate their code and build new nodes forredistribution.

The Time Series node is a method of investigating time series data. It belongs to the Modify category of theSAS SEMMA (sample, explore, modify, model, assess) data mining process. The Time Series node enablesyou to understand trends and seasonal variation in large amounts of time series and transactional data.

The Time Series node in SAS Enterprise Miner enables you to do the following:

� perform time series analysis

� perform forecasting

� work with transactional data

References

Amal, S., and Weselowski, R. (1993). “Practical Econometric Analysis for Assessment ofReal Property Using the SAS System on Personal Computers.” In Proceedings of the Eigh-teenth Annual SAS Users Group International Conference, 385–390. Cary, NC: SAS InstituteInc. https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-62%20Amal%20Weselowski.pdf.

Benseman, B. R. (1990). “Better Forecasting with SAS/ETS Software.” In Proceedings of theFifteenth Annual SAS Users Group International Conference, 494–497. Cary, NC: SAS InstituteInc. https://support.sas.com/resources/papers/proceedings-archive/SUGI90/Sugi-90-74%20Benseman.pdf.

Calise, A., and Earley, J. (1997). “Forecasting College Enrollment Using the SAS System.” In Proceedingsof the Twenty-Second Annual SAS Users Group International Conference, 1326–1329. Cary, NC: SASInstitute Inc. http://www2.sas.com/proceedings/sugi22/STATS/PAPER282.PDF.

Earley, J., Sweeney, S. J., and Zekavat, S. M. (1989). “SAS Goes to Wall Street: PROC ARIMA and theDow Jones Stock Index.” In Proceedings of the Fourteenth Annual SAS Users Group International Confer-ence, 371–375. Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI89/Sugi-89-57%20Earley%20Sweeney%20Zekavat.pdf.

Fischetti, T., Heathcote, S., and Perry, D. (1993). “Using SAS to Create a Modular Forecasting System.” InProceedings of the Eighteenth Annual SAS Users Group International Conference, 580–585. Cary, NC:SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-93%20Fischetti%20Heathcote%20Perry.pdf.

https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-62%20Amal%20Weselowski.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-62%20Amal%20Weselowski.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI90/Sugi-90-74%20Benseman.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI90/Sugi-90-74%20Benseman.pdf

http://www2.sas.com/proceedings/sugi22/STATS/PAPER282.PDF

https://support.sas.com/resources/papers/proceedings-archive/SUGI89/Sugi-89-57%20Earley%20Sweeney%20Zekavat.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI89/Sugi-89-57%20Earley%20Sweeney%20Zekavat.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-93%20Fischetti%20Heathcote%20Perry.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-93%20Fischetti%20Heathcote%20Perry.pdf


Fleming, N. S., Gibson, E., and Fleming, D. G. (1996). “The Use of PROC ARIMA to Test an InterventionEffect.” In Proceedings of the Twenty-First Annual SAS Users Group International Conference, 1317–1326.Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI96/Sugi-96-220%20Fleming%20Gibson%20Fleming.pdf.

Hisnanick, J. J. (1991). “SAS/ETS in Applied Econometrics: Evaluating Input Separability in a Model ofthe U.S. Manufacturing Sector.” In Proceedings of the Sixteenth Annual SAS Users Group InternationalConference, 688–693. Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI91/Sugi-91-116%20Hisnanick.pdf.

Hisnanick, J. J. (1992). “Using PROC ARIMA in Forecasting the Demand and Utilization of InpatientHospital Services.” In Proceedings of the Seventeenth Annual SAS Users Group International Confer-ence, 383–391. Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI92/Sugi-92-64%20Hisnanick.pdf.

Hisnanick, J. J. (1993). “Using SAS/ETS in Applied Econometrics: Parameter Estimates for the CES-Translog Specification.” In Proceedings of the Eighteenth Annual SAS Users Group International Confer-ence, 275–279. Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-46%20Hisnanick.pdf.

Hoyer, K. K., and Gross, K. C. (1993). “Spectral Decomposition and Reconstruction of Nuclear PlantSignals.” In Proceedings of the Eighteenth Annual SAS Users Group International Conference, 1153–1158.Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-193%20Hoyer%20Gross.pdf.

Keshani, D. A., and Taylor, T. N. (1992). “Weather Sensitive Appliance Load Curves; Conditional Demand Es-timation.” In Proceedings of the Seventeenth Annual SAS Users Group International Conference, 422–430.Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI92/Sugi-92-69%20Keshani%20Taylor.pdf.

Khan, M. H. (1990). “Transfer Function Model for Gloss Prediction of Coated Aluminum Using the ARIMAProcedure.” In Proceedings of the Fifteenth Annual SAS Users Group International Conference, 517–522.Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI90/Sugi-90-77%20Khan.pdf.

LaBarr, A. (2014). “Volatility Estimation through ARCH/GARCH Modeling.” In Proceedings of theSAS Global Forum 2014 Conference. Cary, NC: SAS Institute Inc. http://support.sas.com/resources/papers/proceedings14/1456-2014.pdf.

LeBouton, K. J. (1989). “Performance Function for Aircraft Production Using PROC SYSLIN and L2 NormEstimation.” In Proceedings of the Fourteenth Annual SAS Users Group International Conference, 424–426.Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI89/Sugi-89-66%20LeBouton.pdf.

Lin, L.-H., and Myers, S. C. (1988). “Forecasting the Economy Using the Composite Leading Index, ItsComponents, and a Rational Expectations Alternative.” In Proceedings of the Thirteenth Annual SAS UsersGroup International Conference, 181–186. Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI88/Sugi-13-34%20Lin%20Myers.pdf.

McCarty, L. (1994). “Forecasting Operational Indices Using SAS/ETS Software.” In Proceed-ings of the Nineteenth Annual SAS Users Group International Conference, 844–848. Cary, NC:

https://support.sas.com/resources/papers/proceedings-archive/SUGI96/Sugi-96-220%20Fleming%20Gibson%20Fleming.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI96/Sugi-96-220%20Fleming%20Gibson%20Fleming.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI91/Sugi-91-116%20Hisnanick.pdf






https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-193%20Hoyer%20Gross.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI93/Sugi-93-193%20Hoyer%20Gross.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI92/Sugi-92-69%20Keshani%20Taylor.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI92/Sugi-92-69%20Keshani%20Taylor.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI90/Sugi-90-77%20Khan.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI90/Sugi-90-77%20Khan.pdf

http://support.sas.com/resources/papers/proceedings14/1456-2014.pdf


https://support.sas.com/resources/papers/proceedings-archive/SUGI89/Sugi-89-66%20LeBouton.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI89/Sugi-89-66%20LeBouton.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI88/Sugi-13-34%20Lin%20Myers.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI88/Sugi-13-34%20Lin%20Myers.pdf

References F 61

SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI94/Sugi-94-147%20McCarty.pdf.

McNelis, P. D., and Nickelsburg, J. (2002). “Neural Networks and Genetic Algorithms as Tools forForecasting Demand in Consumer Durables (Automobiles).” In Proceedings of the Twenty-Seventh AnnualSAS Users Group International Conference. Cary, NC: SAS Institute Inc. http://www2.sas.com/proceedings/sugi27/p245-27.pdf.

Milhøj, A. (2012). “Analyzing the Time Series of U.S. E-Commerce Using PROC UCM.” In Proceedings ofthe SAS Global Forum 2012 Conference. Cary, NC: SAS Institute Inc. http://support.sas.com/resources/papers/proceedings12/339-2012.pdf.

Milhøj, A. (2015). “Analyzing Parking Meter Transactions Using SAS® Procedures.” In Proceedings ofthe SAS Global Forum 2015 Conference. Cary, NC: SAS Institute Inc. http://support.sas.com/resources/papers/proceedings15/3351-2015.pdf.

Morelock, M. M., Pargellis, C. A., Graham, E. T., Lamarre, D., and Jung, G. (1995). “Time-ResolvedLigand Exchange Reactions: Kinetic Models for Competitive Inhibitors with Recombinant Human Renin.”Journal of Medical Chemistry 38:1751–1761.

Parresol, B. R., and Thomas, C. E. (1991). “Econometric Modeling of Sweetgum Stem Biomass Using theIML and SYSLIN Procedures.” In Proceedings of the Sixteenth Annual SAS Users Group InternationalConference, 694–699. Cary, NC: SAS Institute Inc. https://support.sas.com/resources/papers/proceedings-archive/SUGI91/Sugi-91-117%20Parresol%20Thomas.pdf.

Ragavan, A. J., and Fernandez, G. C. (2006). “Modeling Water Quality Trend in Long Term Time Series.”In Proceedings of the Thirty-First Annual SAS Users Group International Conference. Cary, NC: SASInstitute Inc. http://www2.sas.com/proceedings/sugi31/205-31.pdf.

Rozo, B. J. G., Crook, J., and Moreira, F. (2015). “Modelling Operational Risk Using Extreme ValueTheory and Skew t-Copulas via Bayesian Inference Using SAS.” In Proceedings of the SAS Global Forum2015 Conference. Cary, NC: SAS Institute Inc. http://support.sas.com/resources/papers/proceedings15/3359-2015.pdf.

Shtatland, E. S., and Shtatland, T. (2008). “Another Look at Low-Order Autoregressive Models in EarlyDetection of Epidemic Outbreaks and Explosive Behaviors in Economic and Financial Time Series.” InProceedings of the SAS Global Forum 2008 Conference. Cary, NC: SAS Institute Inc. http://www2.sas.com/proceedings/forum2008/363-2008.pdf.

Wongdhamma, W. (2016). “Upgrade from ARIMA to ARIMAX to Improve Forecasting Accuracy ofNonlinear Time-Series: Create Your Own Exogenous Variables Using Wavelet Analysis.” In Proceedingsof the SAS Global Forum 2016 Conference. Cary, NC: SAS Institute Inc. http://support.sas.com/resources/papers/proceedings16/11823-2016.pdf.

https://support.sas.com/resources/papers/proceedings-archive/SUGI94/Sugi-94-147%20McCarty.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI94/Sugi-94-147%20McCarty.pdf

http://www2.sas.com/proceedings/sugi27/p245-27.pdf

http://www2.sas.com/proceedings/sugi27/p245-27.pdf





https://support.sas.com/resources/papers/proceedings-archive/SUGI91/Sugi-91-117%20Parresol%20Thomas.pdf

https://support.sas.com/resources/papers/proceedings-archive/SUGI91/Sugi-91-117%20Parresol%20Thomas.pdf

http://www2.sas.com/proceedings/sugi31/205-31.pdf



http://www2.sas.com/proceedings/forum2008/363-2008.pdf

http://www2.sas.com/proceedings/forum2008/363-2008.pdf



Index

AMO, 58

Base SAS software, 50

CATALOG procedure, 50SAS catalogs, 50

character functions, 52CHART procedure, 50

histograms, 50COMPARE procedure, 50

comparing SAS data sets, 50comparing SAS data sets, see COMPARE procedureCONTENTS procedure, 50control charts, 55COPY procedure, 50copying SAS data sets, see COPY procedureCORR procedure, 50CPORT procedure, 50crosstabulations, see FREQ procedure

DATA step, 50SAS data sets, 50

DATASETS procedure, 50descriptive statistics, see UNIVARIATE procedure

econometricsfeatures in SAS/ETS software, 16

experimental design, 55

FCMP procedure, 51SAS functions, 51

features in SAS/ETS softwareeconometrics, 16

financial functions, 52FREQ procedure, 51

crosstabulations, 51functions, 52

global statements, 51

histograms, see CHART procedure

IML, see SAS/IML software

JMP, 56

mathematical functions, 52matrix language

SAS/IML software, 54maximizing likelihood functions, 56

MEANS procedure, 51menu interfaces

to SAS/ETS software, 48moving between computer systems

SAS data sets, 50

operations researchSAS/OR software, 55

order statistics, see RANK procedure

Pareto charts, 55PLOT procedure, 51PRINT procedure, 51

printing SAS data sets, 51printing SAS data sets, see PRINT procedureprobability functions, 52PROTO procedure, 51

printing SAS data sets, 51

random number functions, 52RANK procedure, 51

order statistics, 51renaming SAS data sets, 50

SAS catalogs, see CATALOG procedureSAS data sets

contents of, 50copying, 50DATA step, 50moving between computer systems, 50printing, 51renaming, 50sorting, 51structured query language, 51summarizing, 51transposing, 51

SAS Enterprise Guide, 57SAS Enterprise Miner—Time Series node, 59SAS Forecast Studio, 53SAS/ETS software

menu interfaces to, 48SAS/IML software, 54

IML, 54matrix language, 54

SAS/OR software, 55operations research, 55

SAS/QC software, 55statistical quality control, 55

SAS/STAT software, 53

Shewhart control charts, 55SORT procedure, 51

sorting, 51sorting, see SORT procedure

SAS data sets, 51SQL procedure, 51

structured query language, 51STANDARD procedure, 51

standardized values, 51standardized values, see STANDARD procedurestatistical quality control

SAS/QC software, 55structured query language, see SQL procedure

SAS data sets, 51subsetting data, see WHERE statementsummarizing

SAS data sets, 51summarizing SAS data sets, 51

TABULATE procedure, 51tabulating data, 51

tabulating data, see TABULATE procedureTime Series Forecasting menu system, 48TIMEPLOT procedure, 51TRANSPOSE procedure, 51

transposing SAS data sets, 51transposing SAS data sets, see TRANSPOSE

procedure

UNIVARIATE procedure, 51descriptive statistics, 51

WHERE statementsubsetting data, 52

sas/ets 15.2 user’s guide

Documents