Advanced Predictive Modeling Using SAS Enterprise Miner 5

Duration: 2.0 days

This course teaches you how to optimize the performance of predictive models beyond the basics. The course continues the development of predictive models that begins in the Applied Analytics Using SAS Enterprise Miner 5 course.

Learn how to

• use advanced techniques for input selection and model assessment
• construct and evaluate two-stage and multi-stage models using SAS Enterprise Miner
• evaluate variability in model predictive performance.

Who should attend: Predictive modelers and data analysts

Prerequisites

Before attending this course, you should

• have completed the Predictive Modeling Using SAS Enterprise Miner 5.1 or Applied Analytics Using SAS Enterprise Miner 5 course
• have some experience with creating and managing SAS data sets, which you can gain from the SAS Programming I: Essentials course
• have some experience building statistical models using SAS/STAT software
• have completed a statistics course that covers linear regression and logistic regression, such as the Statistics I: Introduction to ANOVA, Regression, and Logistic Regression course.

Course Contents

Review of Basic Predictive Modeling Techniques

• creating a predictive model using SAS Enterprise Miner
• analytic challenges

Improving Input Selection

• univariate screening
• principal components
• variable clusters
• categorical input recoding
• all-subsets regression

Empirical Logits and Model Adequacy

• empirical logit plots
• input transformations

Generalized Profit Assessment

• case-dependent profits
• generalized profit plots
• total profit fraction plots

Building and Evaluating a Two-Stage Model

• assessing models without a profit matrix
• building an interval target model
• non-normal error distributions
• regression trees
• interval target neural network models

Prediction Limits

• profit variability
• generalized profit plots with prediction limits

Software Addressed

This course addresses the following software product(s): SAS Enterprise Miner.

This page was created using SAS software.