Handbook of Statistical Analysis and Data Mining Applications

Chapter 13

Model Evaluation and Enhancement

OUTLINE

Preamble

Introduction

Model Evaluation

Re-Cap of the Most Popular Algorithms

Enhancement Action Checklist

Ensembles of Models: The SingleGreatest Enhancement Technique

How to Thrive as a Data Miner

Postscript

Preamble

One of the most common questions asked by beginning data miners is “How do I know when my model is any good?” This chapter will introduce you to a number of model metrics that you can use to measure the “goodness” of your model. But this process of modeling and evaluation is not a linear process; it is iterative. It is very rare that the best model will be trained initially. Often, the evaluation process will point out some issues that can be resolved by making some changes ...