ABSTRACT
Multivariate predictive modeling is a big data analytics technique that applies merged, integrated data sets to the task of fitting an outcome variable. Classification models are natural extensions of a segmentation project, because a marketer may aim to predict a prospect’s most likely segment, before offering terms of a relationship. Another special type of classification model is binary classification, which predicts a True/False relationship. Predictive modeling is tailor-made for sales and marketing applications, as they can be applied to the full customer journey. Databases for sales and marketing applications are usually set up in a relational structure, with dimension tables and fact tables. This is recommended for storage efficiency, modularity, and maintenance. The chapter provides a general overview of a range of the most commonly employed predictive modeling techniques within sales and marketing. A neural network is a predictive model that fits non-linear relationships and is based on weighted connections between nodes, similar conceptually to neurons in the brain.
