A Comparative Analysis of Feature Selection Methods and Associated Machine Learning Algorithms on Wisconsin Breast Cancer Dataset (WBCD)

Abstract

The sector that affects all the citizens as a common factor is the health sector. It has huge data and important information about patients and their health. Now it is the need of the hour to utilize those enormous data and to extract knowledge, which will be useful to all the stake holders related to electronic health records. In this paper, we have focused on feature selection techniques to gain high quality attributes to enhance the mining process. Feature selection techniques touch all sectors, which require knowledge discovery. In this study, we have made a comparison between different feature selection methods and associated machine learning algorithms on Wisconsin Breast Cancer dataset, Wisconsin Diagnostic Breast Cancer (WDBC), and Wisconsin Prognostic Breast Cancer (WPBC) of UCI Repository. The study found that fusion of classification algorithms perform better on WBCD. It also revealed that IG performs better on WBCD, whereas IG and CFS give good results on WPBC, and CA gives best results on WDBC.