May 2009
Beginner to intermediate
864 pages
23h 13m
English
Robert Nisbet, Ph.D.
OUTLINE
In this exercise, we will use a generalized approach for creating customer response models. This tutorial will analyze churn patterns in a real telecommunications data set. Some fields of the data set have been recoded to preserve anonymity. The Temporal Abstracts have already been calculated with SQL in the database pivot process (see Chapter 4 for a discussion of temporal abstracts). The reference date is the date that records in the training set churned (disconnected service). This pivoting process operates directly on the database early in the Data Preparation phase of modeling. Your job is to select the variables that produce a model with a lift ...