Feature Selection Facilitated Classification For Breast Cancer Prediction

Abstract

Breast cancer is emerging as a torrid research area which attacks the women at an unprecedented rate. In this research we have concentrated on prediction of the breast cancer with few attributes. We have employed feature selection as the preprocessing step for the classification. We used three classifiers and two feature selection strategies for this paper. This work is mainly focused on using the minimal number of attributes for the prediction of cancer in order to reduce the data handling overhead. The preprocessing of data by feature selection creates a major impact in the outcome parameters of the classifier. The results are presented with the detailed discussion.