time series analysis -- an introduction -- ams 586zhu/ams586/time_series_introduction.pdf · time...

38
Time Series Analysis -- An Introduction -- 1 AMS 586

Upload: others

Post on 31-May-2020

13 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Time Series Analysis -- An Introduction --

1

AMS 586

Page 2: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Objectives of time series analysis

Data description

Data interpretation

Modeling

Control

Prediction & Forecasting

2

Page 3: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Time-Series Data

• Numerical data obtained at regular time intervals

• The time intervals can be annually, quarterly, monthly, weekly, daily, hourly, etc.

• Example:

Year: 2005 2006 2007 2008 2009

Sales: 75.3 74.2 78.5 79.7 80.2

3

Page 4: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Time Plot

• the vertical axis measures the variable of interest

• the horizontal axis corresponds to the time periods

U.S. Inflation Rate

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

19

75

19

77

19

79

19

81

19

83

19

85

19

87

19

89

19

91

19

93

19

95

19

97

19

99

20

01

Year

Infl

ati

on

Rate

(%

)

A time-series plot (time plot) is a two-dimensional plot of time series data

4

Page 5: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Time-Series Components

Time Series

Cyclical Component

Irregular /Random Component

Trend Component Seasonal Component

Overall, persistent, long-term movement

Regular periodic fluctuations,

usually within a 12-month period

Repeating swings or movements over more than

one year

Erratic or residual fluctuations

5

Page 6: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Trend Component

• Long-run increase or decrease over time (overall upward or downward movement)

• Data taken over a long period of time

Sales

Time 6

Page 7: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Downward linear trend

Trend Component

• Trend can be upward or downward

• Trend can be linear or non-linear

Sales

Time Upward nonlinear trend

Sales

Time

(continued)

7

Page 8: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Seasonal Component

• Short-term regular wave-like patterns • Observed within 1 year • Often monthly or quarterly

Sales

Time (Quarterly)

Winter

Spring

Summer

Fall

Winter

Spring

Summer

Fall

8

Page 9: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Cyclical Component

• Long-term wave-like patterns

• Regularly occur but may vary in length

• Often measured peak to peak or trough to trough

Sales

1 Cycle

Year 9

Page 10: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Irregular/Random Component

• Unpredictable, random, “residual” fluctuations

• “Noise” in the time series

• The truly irregular component may not be estimated – however, the more predictable random component can be estimated – and is usually the emphasis of time series analysis via the usual stationary time series models such as AR, MA, ARMA etc after we filter out the trend, seasonal and other cyclical components

10

Page 11: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Two simplified time series models

• In the following, we present two classes of simplified time series models 1. Non-seasonal Model with Trend 2. Classical Decomposition Model with Trend and

Seasonal Components

• The usual procedure is to first filter out the trend and seasonal component – then fit the random component with a stationary time series model to capture the correlation structure in the time series

• If necessary, the entire time series (with seasonal, trend, and random components) can be re-analyzed for better estimation, modeling and prediction.

11

Page 12: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

odelsMeasonal s-Non with Trend

trend Stochastic process

random noise

Xt = mt + Yt

12

Page 13: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

odelMecomposition DClassical with Trend and Season

seasonal component

trend Stochastic process

random noise

Xt = mt + st + Yt

13

Page 14: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

odelsMeasonal s-Non with Trend

14

There are two basic methods for estimating/eliminating trend:

Method 1: Trend estimation (first we estimate the trend either by moving average smoothing or regression analysis – then we remove it)

Method 2: Trend elimination by differencing

Page 15: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Method 1: Trend Estimation by Regression Analysis

Estimate a trend line using regression analysis

Year

Time Period

(t)

Sales (X)

2004

2005

2006

2007

2008

2009

0

1

2

3

4

5

20

40

30

50

70

65

Use time (t) as the independent variable:

In least squares linear, non-linear, and exponential modeling, time periods are numbered starting with 0 and increasing by 1 for each time period.

15

Page 16: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Least Squares Regression

The estimated linear trend equation is:

Sales trend

0

10

20

30

40

50

60

70

80

0 1 2 3 4 5 6

Year

sale

s

Year

Time

Period

(t)

Sales

(X)

2004

2005

2006

2007

2008

2009

0

1

2

3

4

5

20

40

30

50

70

65

Without knowing the exact time series random error correlation structure, one often resorts to the ordinary least squares regression method, not optimal but practical.

16

Page 17: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Linear Trend Forecasting

• Forecast for time period 6 (2010):

Year

Time

Period

(t)

Sales

(X)

2004

2005

2006

2007

2008

2009

2010

0

1

2

3

4

5

6

20

40

30

50

70

65

??

One can even performs trend forecasting at this point – but bear in mind that the forecasting may not be optimal.

Sales trend

0

10

20

30

40

50

60

70

80

0 1 2 3 4 5 6

Year

sale

s

17

Page 18: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Method 2: Trend Elimination by Differencing

18

Page 19: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Trend Elimination by Differencing

19

If the operator ∇ is applied to a linear trend function:

Then we obtain the constant function:

In the same way any polynomial trend of degree k can be removed by the operator:

Page 20: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

odelMecomposition DClassical (Seasonal Model)

with trend and season

20

where

Page 21: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

odel Mecomposition DClassical

21

we remove the seasonal First Differencing:: 2Method component by differencing. We then remove the trend by differencing as well.

can fit a Alternatively, wefit method: -Joint: 3Method combined polynomial linear regression and harmonic functions to estimate and then remove the trend and seasonal component simultaneously as the following:

we estimate and remove the First Filtering:: 1Method trend component by using moving average method; then we estimate and remove the seasonal component by using suitable periodic averages.

Page 22: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Method 1: Filtering

22

(1). We first estimate the trend by the moving average: • If d = 2q (even), we use:

• If d = 2q+1 (odd), we use:

(2). Then we estimate the seasonal component by using the average

, k = 1, …, d, of the de-trended data:

To ensure:

we further subtract the mean of

(3). One can also re-analyze the trend from the de-seasonalized data in order to obtain a polynomial linear regression equation for modeling and prediction purposes.

Page 23: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Method 2: Differencing

23

Define the lag-d differencing operator as:

We can transform a seasonal model to a non-seasonal model:

Differencing method can then be further applied to eliminate the trend component.

Page 24: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Method 3: Joint Modeling

24

As shown before, one can also fit a joint model to analyze both components simultaneously:

Page 25: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Detrended series

P. J. Brockwell, R. A. Davis, Introduction to Time Series and Forecasting, Springer, 1987 25

Page 26: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Time series – Realization of a stochastic process

{Xt} is a stochastic time series if each component takes a value according to a certain probability

distribution function.

A time series model specifies the joint distribution of the sequence of random variables.

26

Page 27: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

White noise - example of a time series model

27

Page 28: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Gaussian white noise

28

Page 29: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Stochastic properties of the process

STATIONARITY

.1Once we have removed the seasonal and trend components of a time series (as in the classical decomposition model), the remainder (random) component – the residual, can often be modeled by a stationary time series.

* System does not change its properties in time

* Well-developed analytical methods of signal analysis and stochastic processes

29

Page 30: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

WHEN A STOCHASTIC PROCESS IS STATIONARY?

{Xt} is a strictly stationary time series if

f(X1,...,Xn)= f(X1+h,...,Xn+h),

where n1, h – integer

Properties:

* The random variables are identically distributed.

* An idependent identically distributed (iid) sequence is strictly stationary.

30

Page 31: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Weak stationarity

{Xt} is a weakly stationary time series if

• EXt = and Var(Xt) = 2 are independent of time t

• Cov(Xs, Xr) depends on (s-r) only, independent of t

31

Page 32: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Autocorrelation function (ACF)

32

Page 33: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

ACF for Gaussian WN

33

Page 34: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

ARMA models

Time series is an ARMA(p,q) process if Xt is stationary and if for every t:

Xt 1Xt-1 ... pXt-p= Zt + 1Zt-1 +...+ pZt-p where Zt represents white noise with mean 0 and variance 2

The Left side of the equation represents the Autoregressive AR(p) part, and the right side the Moving Average MA(q) component.

34

Page 35: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Examples

35

Page 36: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Exponential decay of ACF

MA(1) sample ACF

AR(1)

36

Page 38: Time Series Analysis -- An Introduction -- AMS 586zhu/ams586/Time_Series_Introduction.pdf · Time Series: Theory and Methods. Springer-Verlag. Brockwell, Peter J. and Davis, Richard

Reference

38

Box, George and Jenkins, Gwilym (1970) Time series analysis: Forecasting and control, San Francisco: Holden-Day. Brockwell, Peter J. and Davis, Richard A. (1991). Time Series: Theory and Methods. Springer-Verlag.

Brockwell, Peter J. and Davis, Richard A. (1987, 2002). Introduction to Time Series and Forecasting. Springer.

We also thank various on-line open resources for time series analysis.