w e l c o m e interest group big data analytics
TRANSCRIPT
W E L C O M EInterest Group Big Data Analytics
Co-chairs Morris Riedel, Kuo Kwo-Sen, Peter Baumann
Scientific Big Data Analytics
Classification++
Regression++Clustering++
Serial Algorithms for Large Volumes of Data Exist ‘Big Data‘ Requires Parallel Algorithms and Open Availability
MLlib
Scientific Big Data Analytics
„Reference Data Analytics“for reusability & learning
CRISP-DM
Report
OpenlyShared
Datasets
RunningAnalytics
Code
Big Data Analytics IGBig Data Infrastructure WG
Towards Systematic Data Analytics
Guided by the Cross Industry Standard Process for Data Mining (CRISP-DM) Phases
‘Building a UCI Repository for Big Data Analytics‘
Scientific Big Data Analytics
Sattelite Data
Parallel Support VectorMachines (SVM)
HPC/MPI, Map-Reduce & GPGPUs
ClassificationStudy of
Land CoverTypes
„Reference Data Analytics“for reusability & learning
CRISP-DM
Report
OpenlyShared
Datasets
RunningAnalytics
Code
‘Best Practices‘
Community-based practice
ParallelBrain DataAnalytics
(Quickbird)
Big Data Analytics IGBig Data Infrastructure WG
Classification++
G. Cavallaro and M. Riedel, ‘Smart Data Analytics Methods for Remote Sensing Applications‘, IGARSS 2014
Results Example