BUS5PA Assignment 2: Predictive Analysis of Organic Product Purchases

This report presents an analysis of a predictive modeling assignment (BUS5PA Assignment 2) using SAS Enterprise Miner. The objective was to predict which customer segments are likely to purchase a new line of organic products. The project involved setting up the project, performing exploratory data analysis, and building and evaluating predictive models using Decision Tree and Regression modeling techniques. The report details the roles and measurement levels of the variables, distribution of target variables, and the rejection of certain variables based on their characteristics. The Decision Tree modeling section describes the creation of two Decision Tree models, including the selection of optimal trees based on Average Square Error (ASE), the use of Logworth for variable splitting, and the interpretation of the rules generated by each tree. The Regression modeling section covers data imputation, variable clustering, the selection of a Stepwise model, and the application of Logistic Regression due to the binary nature of the target variable. The report outlines the variables included in the final Regression model, their significance, and the interpretation of their coefficients. Furthermore, a model comparison tool was used to assess the performance of the models and determine which one outperformed the other. The report also includes open-ended discussion about the findings and their implications.