Clustering and Association Modeling Using IBM SPSS Modeler (v18.1.1)

Audience

• Modelers, Analysts

Prerequisites

• Experience using IBM SPSS Modeler
• A familiarity with the IBM SPSS Modeler environment: creating models, creating streams, reading in data files, and assessing data quality
• A familiarity with handling missing data (including Type and Data Audit nodes), and basic data manipulation (including Derive and Select nodes)

Course topics

1: Introduction to clustering and association modeling

• Identify the association and clustering modeling techniques available in IBM SPSS Modeler
• Explore the association and clustering modeling techniques available in IBM SPSS Modeler
• Discuss when to use a particular technique on what type of data

2: Clustering models and K-Means clustering

• Identify basic clustering models in IBM SPSS Modeler
• Identify the basic characteristics of cluster analysis
• Recognize cluster validation techniques
• Understand K-Means clustering principles
• Identify the configuration of the K-means node

3: Clustering using the Kohonen network

• Identify the basic characteristics of the Kohonen network
• Understand how to configure a Kohonen node
• Model a Kohonen network

4: Clustering using TwoStep clustering

• Identify the basic characteristics of TwoStep clustering
• Identify the basic characteristics of TwoStep-AS clustering
• Model and analyze a TwoStep clustering solution

5: Use Apriori to generate association rules

• Identify three methods of generating association rules
• Use the Apriori node to build a set of association rules
• Interpret association rules

6: Use advanced options in Apriori

• Identify association modeling terms and rules
• Identify evaluation measures used in association modeling
• Identify the capabilities of the Association Rules node
• Model associations and generate rules using Apriori

7: Sequence detection

• Explore sequence detection association models
• Identify sequence detection methods
• Examine the Sequence node
• Interpret the sequence rules and add sequence predictions to steams

8: Advanced Sequence detection

• Identify advanced sequence detection options used with the Sequence node
• Perform in-depth sequence analysis
• Identify the expert options in the Sequence node
• Search for sequences in Web log data

A: Examine learning rate in Kohonen networks (Optional)

• Understand how a Kohonen neural network learns

B: Association using the Carma model (Optional)

• Review association rules
• Identify the Carma model
• Identify the Carma node
• Model associations and generate rules using Carma