Predictive Analytics: A New Wealth of Options
-
Upload
inside-analysis -
Category
Technology
-
view
912 -
download
1
Transcript of Predictive Analytics: A New Wealth of Options
CONFIDENTIAL
PREDICTIVE ANALYTICS WITH SAP SYBASE IQ JOYDEEP DAS PRODUCT MANAGEMENT
APRIL 17, 2012
© 2012 SAP AG. All rights reserved. 2
BUZZ IN THE INDUSTRY - I Unleashes Business Value
Operational Efficiencies
Revenue Growth
New Strategies & Business Models
Business Value*
© 2012 SAP AG. All rights reserved. 3
BUZZ IN THE INDUSTRY - II New Demands, New Opportunities
Mining New Sources of Data • Large volumes of data beyond structured
• Text, social, clickstream, geospatial • Interactive, on-the-fly analysis • New methodologies
• e.g. social network analysis
Proliferation • Increased usage by non-technical
business users • Embedded in applications
• e.g. recommendation engines in CRM
• Platform has to be accessible and support more users!
© 2012 SAP AG. All rights reserved. 4
SAP Sybase IQ A Powerful Platform For Predictive Analytics – Sample Case Study #1
Sybase IQ significantly enhanced AOK Hessen’s ability to handle complex business predic0ons involving mul0-‐dimensional analysis of many input variables.
AOK Hessen, a large health insurance company, searches for pa+erns across all exis8ng informa8on – medical treatment, prescrip8on, insurance benefits – simultaneously.
“The divisions currently using the tool run a significantly greater number of analyses than ever before. They keep discovering new ways of drilling down into data while working with the soCware.”
-‐ Michael Shimmelpfennig, Service Manager, IT-‐Business Department
© 2012 SAP AG. All rights reserved. 5
SAP Sybase IQ A Powerful Platform For Predictive Analytics – Sample Case Study #2
Sybase IQ enables Playphone to gain significant compe00ve advantage through new capabili>es in customer targe0ng, opera0onal efficiency and fraud detec0on.
Playphone, a leading global mobile entertainment and media company, enables customer analy8cs for large scale marke;ng campaigns using advanced predic8ve models on Sybase IQ -‐ crunching through
customer behavior, purchasing history, and many other relevant metrics.
“Sybase IQ is a brilliant analy;cs engine. I don’t know that the business would s8ll be here today without our Sybase solu8ons.”
-‐ Simon Rose, Director Of Infrastructure
© 2012 SAP AG. All rights reserved. 6
SAP Sybase IQ Our Objective
Find buried signal in time! Or, get buried in Data!
Model Diversity Big Data Complex Alternatives
Deal With It
© 2012 SAP AG. All rights reserved. 7
SAP Sybase IQ PlexQ Technology: Versatile With Multiple Options For Predictive Analytics
Pull Out Push Down
Data Mining Tools
BI/DM Tools
Federated Push - Pull
BI Tools
Accelerated Access Drivers
Embedded libraries
Method I Method II Method III
Full Mesh High Speed Interconnect
IBM SPSS
SAS R
Panopticon
KXEN
Hadoop
R
Fuzzy Logix Visual Numerics
© 2012 SAP AG. All rights reserved. 8
SAP Sybase IQ Method I: Traditional Pull To Client
Pull Out
Data Mining Tools
Accelerated Access Drivers
Method I
Full Mesh High Speed Interconnect
IBM SPSS
SAS R
Standard Drivers Optimized Drivers
Pervasive Method
Myriad Tools
Available Skillset
© 2012 SAP AG. All rights reserved. 9
SAP Sybase IQ
Fetch data from database
Create datasets for analy>cal packages
Analyze data using sta>s>cal func>ons on proprietary plaMorms
Store results from datasets back into database
Generate reports
Time consuming process Could run into memory constraints with large data sets
Proprietary plaMorms make it very difficult to embed in applica8ons
Another 8me consuming process which could slow down the delivery of results to end-‐users
Data
Volume
Processing Time
Accuracy
Compromise on at least one key area
Visualization Database Server Logic/Filtering Applied Outside Database Server
Visualization Logic/Filtering Applied Inside Database Server
!
Overcoming Method I: Avoid Pull
© 2012 SAP AG. All rights reserved. 10
SAP Sybase IQ
Models per month
2000
2005
2010 2015
Factory Analysis
Craftsman Analysis
Fully automated modeling process
• Regression
• Classification
• Segmentation
• Time series forecasting
• Association rules
Identify key variables
Generate Sybase IQ specific SQL code
Executive and operational reports
Easy to Use Time to Market More Models
In-database Push down
KXEN
Push down (Method II): Solution #1 with KXEN
© 2012 SAP AG. All rights reserved. 11
SAP Sybase IQ Push down (Method II): Solution #2 with Fuzzy Logix
Visualization Fuzzy Logix Libraries Executed Inside Sybase IQ
Application Algorithms Direct Marke0ng: Op>mize the performance of Internet Marke>ng Customer Reten>on
K-‐ Means, Correla>on, PorKolio Op>miza>on, PCA, Logis>c Regression
Marke0ng Services: Accelerate the speed of model development for their clients
Sparse Matrix Calcula>ons, Correla>on, Euclidean Distance
Health Insurance: Risk Management and Client Management Teams -‐ Scoring models to assess the quality and efficiency of care
Sparse Matrix Calcula>ons, Correla>on, Euclidean Distance
Banking: Risk Management Correla>on, Simula>on, Regression, Cubic Spline, Matrix Opera>ons
© 2012 SAP AG. All rights reserved. 12
SAP Sybase IQ Push down (Method II): Solution #3 with Zementis
Database Server Sybase IQ
Applica>ons
PMML (models) PMML (models) PMML (models)
Zemen>s PMML Preprocessor (convert & validate)
Universal Plug-‐In
SQL
Predic>ons
UDFs
Bridge
PMML
Express Complex Computa0ons In Industry Standard Predic0ve Modeling Markup Language (PMML), Plug In Models Close To data for execu0on
© 2012 SAP AG. All rights reserved. 13
SAP Sybase IQ Push down (Method II): Solution #3 with Zementis (contd)
Delivers a wide range of model types for high performance scoring, including: • Decision Trees for classification and regression • Neural Network Models: Back-Propagation, Radial-Basis Function, and Neural-Gas • Support Vector Machines for regression, binary and multi-class classification • Linear and Logistic Regression (binary and multinomial) • Naïve Bayes Classifiers • General and Generalized Linear Models • Cox Regression Models • Rule Set Models (flat decision trees) • Clustering Models: Distribution-Based, Center-Based, and 2-Step Clustering • Scorecards (point allocation for categorical, continuous and complex attributes) • Association Rules • Also implements data dictionary, missing / invalid values handling and data pre-processing
Handles most SAS models publishable in PMML from SAS Enterprise Miner
© 2012 SAP AG. All rights reserved. 14
SAP Sybase IQ Push down (Method III): Solution #1 Federated Push-Pull With “R”
PUSH-‐PULL FEDERATION:
-‐ UDF Bridge Between Sybase IQ and “R” server
-‐ Fire SQL against Sybase IQ that pushes “R” models embedded in UDFs to “R” server for execu>on
-‐ UDFs pulls results back for combining with rest of SQL query results in Sybase IQ
-‐ Supports most “R” models
C++ UDF “R” Server
SQL Queries
© 2012 SAP AG. All rights reserved. 15
SAP Sybase IQ Push down (Method III): Solution #2 Client Side Push-Pull
Sybase IQ Predictive Analytics
Job Results
$
QUEST Toad for Cloud Databases
Hadoop Hive MR Job Results
• Ideal for bringing together predictive analytics computations from different domains
• Better performance when computation from each domain is pushed down to each branch and then pulled together
© 2012 SAP AG. All rights reserved. 16
SAP Sybase IQ Push down (Method III): Solution #3 Data Federation Push-Pull
• Ideal for combining subsets of HDFS data with Sybase IQ data for operational analytics
• HDFS data not implicitly stored in Sybase IQ: Fetched into Sybase IQ In-memory tables on the fly as part of query fired at Sybase IQ
Predictive Queries
UDF Bridge Hadoop
Distributed File System
© 2012 SAP AG. All rights reserved. 17
SAP Sybase IQ Push down (Method III): Solution #4 Query Federation Push-Pull
Predictive Queries
UDF Bridge Hadoop
MapReduce Jobs
• Ideal for combining subsets of Hadoop MapReduce job results with Sybase IQ data for operational analytics
• Hadoop MapReduce results not implicitly stored in Sybase IQ: Fetched into Sybase IQ In-memory tables on the fly as part of query fired at Sybase IQ
© 2012 SAP AG. All rights reserved. 18
SAP Sybase IQ Push down (Method III): Solution #5 Federated Push-Pull Text Analytics
Data Files
Text Index
(ISYS)
BLO
B
Web
Service
SAP BusinessObjects
Data Services Sybase IQ
1. Limit Textual Corpus 2. Extract Concepts/Entity Relationships 3. Mine Resulting Schemas
iSYS-‐Search Document Filters
Kapow So[ware
FourSquare
Amazon
© 2012 SAP AG. All rights reserved. 19
SAP Sybase IQ
Data Discovery (Data Scien0sts)
Applica>on Modeling (Business Analysts)
Reports/Dashboards (BI Programmers)
Business Decisions (Business End Users)
Infrastructure Management
(DBAs)
Full Mesh High Speed Interconnect
A Versatile Platform For Predictive Analytics
© 2012 SAP AG. All rights reserved. 20
SAP Sybase IQ Summary – transform your business
• Predictive Analytics going mainstream
• Many options available
• SAP Sybase IQ has a broad and comprehensive support
THANK YOU!
© 2012 SAP AG. All rights reserved. 21
No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of SAP AG. The information contained herein may be changed without prior notice.
Some software products marketed by SAP AG and its distributors contain proprietary software components of other software vendors.
Microsoft, Windows, Excel, Outlook, and PowerPoint are registered trademarks of Microsoft Corporation.
IBM, DB2, DB2 Universal Database, System i, System i5, System p, System p5, System x, System z, System z10, System z9, z10, z9, iSeries, pSeries, xSeries, zSeries, eServer, z/VM, z/OS, i5/OS, S/390, OS/390, OS/400, AS/400, S/390 Parallel Enterprise Server, PowerVM, Power Architecture, POWER6+, POWER6, POWER5+, POWER5, POWER, OpenPower, PowerPC, BatchPipes, BladeCenter, System Storage, GPFS, HACMP, RETAIN, DB2 Connect, RACF, Redbooks, OS/2, Parallel Sysplex, MVS/ESA, AIX, Intelligent Miner, WebSphere, Netfinity, Tivoli and Informix are trademarks or registered trademarks of IBM Corporation.
Linux is the registered trademark of Linus Torvalds in the U.S. and other countries.
Adobe, the Adobe logo, Acrobat, PostScript, and Reader are either trademarks or registered trademarks of Adobe Systems Incorporated in the United States and/or other countries.
Oracle and Java are registered trademarks of Oracle.
UNIX, X/Open, OSF/1, and Motif are registered trademarks of the Open Group.
Citrix, ICA, Program Neighborhood, MetaFrame, WinFrame, VideoFrame, and MultiWin are trademarks or registered trademarks of Citrix Systems, Inc.
HTML, XML, XHTML and W3C are trademarks or registered trademarks of W3C®, World Wide Web Consortium, Massachusetts Institute of Technology.
© 2012 SAP AG. All rights reserved.
SAP, R/3, SAP NetWeaver, Duet, PartnerEdge, ByDesign, SAP BusinessObjects Explorer, StreamWork, SAP HANA, and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of SAP AG in Germany and other countries. Business Objects and the Business Objects logo, BusinessObjects, Crystal Reports, Crystal Decisions, Web Intelligence, Xcelsius, and other Business Objects products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of Business Objects Software Ltd. Business Objects is an SAP company.
Sybase and Adaptive Server, iAnywhere, Sybase 365, SQL Anywhere, and other Sybase products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of Sybase, Inc. Sybase is an SAP company.
All other product and service names mentioned are the trademarks of their respective companies. Data contained in this document serves informational purposes only. National product specifications may vary.
The information in this document is proprietary to SAP. No part of this document may be reproduced, copied, or transmitted in any form or for any purpose without the express prior written permission of SAP AG.