Application of Bayesian Network Classifiers and Data Envelopment Analysis for Mining Breast Cancer Patterns

Abstract

Data envelopment analysis (DEA) and learning Bayesian networks (LBN) are relatively unexplored approaches for data mining. Recent theoretical research indicates that DEA and LBN can be used for certain data mining applications. We apply DEA and LBN for discovering the breast cancer pattern. We use data from a large hospital in Pennsylvania to discover the breast cancer patterns, and benchmark the performance of DEA and LBN against a popular statistical linear discriminant analysis technique. The results of our experiments indicate that DEA and LBN outperform statistical linear discriminant analysis.