The Hausdorff distance measure for feature selection in learning applications

Abstract

Recent advances in computing technology in terms of speed, cost, as well as access to tremendous amounts of computing power and the ability to process huge amounts of data in reasonable time has spurred increased interest in data mining applications. Machine learning has been one of the methods used in most of these data mining applications. It is widely acknowledged that about 80% of the resources in a majority of data mining applications are spent on cleaning and pre-processing the data. However, there have been relatively few studies on pre-processing data used as input in these data mining systems. In this study, we present a feature selection method based on the Hausdorff distance measure, and evaluate its effectiveness in pre-processing input data for inducing decision trees. The Hausdorff distance measure has been used extensively in computer vision and graphics applications to determine the similarity of patterns. Two real-world financial credit scoring data sets are used to illustrate the performance of the proposed method.