Predictive Analytics For Dummies, 2nd Edition

Chapter 7

Predicting the Future Using Data Classification

IN THIS CHAPTER

Introducing data classification

Making better decisions by using data classification

Getting an overview of widely used data classification algorithms

Using data classification algorithms to predict future behavior of customers

Leveraging basic decision trees to assess business opportunities

Introducing the fundamentals of deep learning

Applying data classification as a way to make business predictions

This chapter introduces an easy and widely used concept that can help you predict the future: data classification, a data-mining technique used to make sense of overwhelmingly large and humanly unmanageable datasets to uncover hidden patterns in the data. That could be one possible first step in the direction of building a predictive model.

Predictive models that use this technique are called classification models, often referred to as classifiers. Behind the scenes of these classifiers is a set of sophisticated machine-learning algorithms that learn from experience — as represented by the historical data — to discover hidden patterns and trends in the data. When these unseen relationships come to light, the model can predict the outcome (a category (class label), or numerical value) based on newly input data.

For instance, most financial institutions that issue loans to individuals or corporate entities use risk modeling to predict whether the loan will be paid back (or, in the case of mortgage loans, paid ...