See it in SPSS Clementine 8.5

1
See it in SPSSClementine 8.5

• Laura Squier
• June 15, 2004

2
Agenda

• Customer Success and ROI
• Data Mining Defined
• The Data Mining Process
• Clementine
• Offerings and Capabilities
• In the news
• New features for Clementine 8.5

3
Leading Organizations to Success

• The median ROI for the projects that
  incorporated predictive technologies was 145,
  compared with a median ROI of 89 for those
  projects that did not.
• Source IDC, Predictive Analytics and ROI
  Lessons from IDCs Financial Impact Study

4
IDC Study on Predictive Analytics

• This ROI included enhancements in
• Technology.
• The amount of money saved on technology or
  technology costs avoided by introducing the
  analytic solution, such as diverting data
  processing to a more cost-effective system
• Productivity
• Efficiency savings due to the reduced amount of
  time and effort required for particular tasks
• Business process enhancement
• All identifiable annual savings that were
  realized due to changes in business process
  supported

5
Applications and Industries
Financial Services
Acquisition and Retention Cross-sell,
Up-sell Fraud detection
Health Care
Telecom
Resource utilization Fraud detectionQuality of
care
Cross-sell, Up-sell Retention
Data mining
Customer loyalty Direct Mail response Market
basket
Up-sell, Cross-sell Direct Mail Loyalty
Gaming Entertainment
Fraud detection Collections Employee/Customer
Satisfaction
Retail/CPG
Public sector
6
Clementine in Finance, Banking and Insurance

• Clementine offers techniques and expertise to
  help you
• Enhance your anti-money laundering and Patriot
  Act efforts
• Know your loyal customers
• Identify which customers are likely to purchase
  which bundles of products
• Select good candidates for a loan or credit

7
Customer Success
Improved customer retention and increased profits
by 10 20 by identifying behaviors likely to
lead to churn
Customer retention strategies increased bottom
line profits 10 20
Identified profiles of individuals interested in
higher yielding investments and reduced direct
mail costs by 30 while garnering 95 of the
campaigns revenue.
8
Clementine in Retail and Consumer Packaged Goods

• Clementine offers techniques and expertise to
  help you
• Discover which customers are good candidates for
  purchasing particular items
• Understand what items are purchased together
• Understanding drivers of product demand
• Improve customer loyalty
• Personalize interactions
• Improve and shorten product development

9
Customer Success
Reduced the costs of excess inventory and ordered
proper quantities of popular items yielding a
3.8 savings over the previous year and improving
customer loyalty
Improved cross-selling and achieved web site
personalization, acquiring a 300 year-on-year
rise in profits in the first month.
Shortened the length of time that it takes to get
a product to market and minimized the need for
animal testing.
10
Clementine in Healthcare and Pharmaceuticals

• Clementine offers techniques and expertise to
  help you
• Minimize fraud and noncompliance
• Identify best practices
• Improve patient satisfaction and loyalty
• Conduct bioinformatics research such as
  understanding the genetic components of disease

11
Customer Success
Isolated likely causes of payment errors and in a
sample detected 50 of incorrect payments
reducing the amount on money lost to payment
errors
Discovered which gene was most important in tumor
growth and identified genes that were previously
unknown to be related to tumor biology.
12
Clementine in Telecommunications

• Clementine offers techniques and expertise to
  help you
• Identify good prospects
• Identify which customers are likely to be the
  most loyal
• Understand which customers are likely to churn
• Take actions that might be to minimize the risk
  of churning

13
Customer Success
British Telecom - Doubled direct mail campaign
response rate and build highly accurate models to
pre-empt churn of highly profitable customers
Orange Communications is able to identify 80 of
churners and can provide them with the
appropriate treatment to retain the high valued
customers.
14
Data Mining Defined

• An interactive process which leverages analysis
  technologies and computing power to find patterns
  in data that help improve performance by
• Reducing costs
• Increasing profits
• Better allocating resources
• Finding unusual behavior

15
In a sense, everyone does data mining

• People use information about the world to draw
  conclusions.
• As the number of observations grows large, people
  are no longer able to easily find these patterns.
  Reliance on computing power becomes necessary to
  draw reliable conclusions.

16
What Does Data Mining Do?

• Data mining uses existing data to
• Predict
• Category membership
• Numeric Value
• Group
• Cluster (group) things together based on their
  characteristics
• Associate
• Find events that occur together, or in a sequence
• Find outliers
• Identify cases that dont follow expected
  behavior

17
Methodology for Data Mining

• CRISP-DM Cross Industry Standard Process for
  Data Mining
• Focus on Business Issues, map to Data Mining
  Tasks
• See www.crisp-dm.org for more information

18
Clementines Support for CRISP-DM

• Supports the entire data mining process process
• Provides highly accurate, broad and deep modeling
  techniques for prediction, classification,
  association and clustering
• Allows user to easily and interactively build
  models
• Provides the ability to deploy the entire data
  mining process immediately
• Leverages your existing database investment
• Allows easy integration of external routines
  (CEMI)
• Supports specific business applications through
  templates (CATs)
• Utilizes highly scalable architecture

19
Clementine

• Clementine is a scalable data mining workbench
  that enables you to quickly develop predictive
  models based on business expertise, and deploy
  the models to improve decision making.

20
Clementine in the News

• Clementine ranks 1 data mining product in
  www.kdnuggets.com reader poll. (06/04)
• Clementine 8.5 Released. (05/04)
• InfoWorld Product Review Clementine 8.1 melds
  BA with BI foundation. (05/04)
• Childrens Memorial Hospital and SPSS are
  recognized by Computerworld Honors Foundation for
  Outstanding Achievement in Medicine in
  researching better treatments for pediatric brain
  tumors. (04/04)
• Clementine Named Finalist in 2004 eWEEK
  Excellence Awards (03/04)
• CIO magazine recognizes Continental Airlines for
  a CIO Enterprise Value Award (02/04)
• SPSS Inc. recognized as a Company to Watch by
  Intelligent Enterprise magazine (12/03)

21
Key Clementine Features

• Scales the entire data mining process
• Leverages your existing data warehouse
  investments

22
Key Clementine Features

• Unparalleled productivity
• Workflow provides complete support of the
  complete CRISP-DM data mining process
• Breadth of techniques for modeling and processing
• Vertical templates for specific help

23
Clementine Features - Data Understanding

• Data Understanding Support
• Graphical and tabular support

24
Clementine Features - Data Preparation

• Data preparation support
• Access data from multiple data sources at one
  time
• Automatically recode data
• Automatically decile/rank data
• Specify missing values
• Set Global Values
• Derive new variables
• Aggregate data
• Look back to historical data
• Fill in missing values using mean, calculation,
  prediction from Clementine or from SPSS
• Factor/PCA Analysis
• Merge and append data
• Parse and structure text
• And more

25
Clementine Features - Modeling

• Supervised Learning Techniques
• Regression algorithms neural networks, CRT,
  linear regression
• Classification algorithms C5.0, logistic
  regression, CRT, Neural Networks

• Unsupervised or quasi-supervised techniques
• Clustering K-means, Kohonen, Two-Step
• Association algorithms apriori, GRI, Sequence

26
Clementine Features - Evaluation

• Multiple ways to evaluate model performance
• Evaluate multiple models at once
• Overall model performance on testing sample
• Compare with business rules or best fit
• Misclassification matrices
• Confidence Values
• Evaluations charts gains, lift, response,
  profit, and ROI

27
Clementine Features - Deployment

• Clementine support various modes of deployment
  depending on user needs
• Deployment Options
• Deploy stream results into other SPSS
  Applications
• Embed scoring into existing business
  infrastructure

28
Key Clementine Features

• Ease of Deployment
• Fast and cost-effective delivery of data mining
  solutions

Model Export C code or XML
Clementine Solution Publisher model and
processing( no need to compile)
29
Clementine Offerings

• Clementine
• SPSS data mining workbench
• Clementine Server
• SPSS data mining server
• Available on Windows Sun Solaris, IBM AIX ,IBM
  OS/400 (iSeries), HP-UX
• Clementine Solution Publisher
• Provides the ability to publish entire data
  mining process to be used on a separate platform

• Cleo
• Provides the ability to publish and deploy the
  entire data mining process to a web environment
• Text Mining for Clementine
• Provides the ability to extract concepts from
  text to use in Clementine
• Web Mining for Clementine
• Provides the ability to extract events from web
  logs for fast web analysis in Clementine
• Includes Web CAT
• Clementine Application Templates
• Quick start templates designed for specific
  application areas
• For CRM, Telecommunications, Fraud, Web Mining,
  and Microarray analysis

30
Whats new in Clementine 8.5?
31
Clementine 8.5 New Features

• Integration
• In-database mining for Microsoft Decision Trees
• Export models (PMML) to be viewed IM
  Visualization
• Metadata support for field level formatting
• Deployment
• Export models (PMML) to be used in IBM DB2 Data
  Warehouse Edition 8.2
• Ability to deploy Clementine models into SPSS
  Predictive Applications (formally DD)
• CARMA Association Model, scoring with association
  models

32
In-Database Mining

• Access, build, and deploy MS Decision Trees
  directly from the Clementine interface
• Works with Microsoft Analysis Services and SQL
  Server

33
Integration with IBM DB2 Data Warehouse Edition

• Browse association models in IM Visualization
• Export models (PMML) for use in IM Scoring
• New web page on integration http//www.spss.com/c
  lementine/ibmdb2.htm

34
Metadata Support

• Clementine maintains metadata information from
  SPSS data
• Variable and Value labels
• Field level formatting
• Ordinal Data support

35
Deployment into SPSS Predictive Applications

• Clementine Wizard packages models
• Models may be used in real time to deploy to all
  customer touch points.

36
New Association Algorithm

• Provides model deployment
• Ability to browse the trellis of the association
  models, filter, sort and select to deploy on
  specific rules.

37
The Future of Clementine

• SPSS data mining strategy for the future is two
  fold
• To continue to make Clementine the best
  horizontal tool
• Leverage more database capabilities
• Enhance visualization
• Add requested algorithms
• Clementine support of the framework to create and
  deliver high value, analytical applications

38
Contact Information

• Please contact sales_at_spss.com
• Or call 1-800-543-2185

39
Questions?

• Laura Squier
• Clementine Product Manager
• lsquier_at_spss.com