Data Mining Algorithms: Explained Using R

Chapter 17

Attribute transformation

17.1 Introduction

Attribute transformation alters the data by replacing a selected attribute by one or more new attributes, functionally dependent on the original one, to facilitate further analysis. This is typically performed prior to creating classification, regression, or clustering models to bypass some limitations of the modeling algorithms used or improve model quality. The latter is possible, since—like attribute selection covered in Chapter 19—attribute transformation modifies the space of models being searched by changing the representation of instances and may make good models more likely or easier to find. This chapter presents the description and examples illustrating the most commonly used attribute transformations.

Ex. 1.3.2 dmr.data

Example 17.1.1

Attribute transformation algorithms presented in this chapter will be illustrated by R code examples, containing simple implementations and usage demonstrations thereof. The latter will be using the toy weatherc data, as well as the more realistic Vehicle Silhouettes and Glass datasets from the mlbench package. The effects of some transformations will be illustrated by creating classification models based on the original and transformed datasets. The dmr.claseval and dmr.util packages will be used to evaluate prediction quality and provide auxiliary utility functions. The popular R implementations of the decision tree and naïve Bayes algorithms, provided by the rpart