A RFE- Optimised Method for Predicting the Malignancy of the Breast Cancer

摘要

Breast cancer has the most mortality rate in the case of cancer after the skin cancer. Early detection of the tumour cells would help in different therapeutic treatments. In our study, RFE (recursive feature elimination) is implemented for selecting the optimal set of features which could help on characterizing the tumor cells. The feature size has been reduced to half making the algorithm faster. The combination of RFE has been implemented with different types of classifiers and the best result was obtained in the case of logistic regression with an accuracy of 96.5% and an auc score of 0.98. This study could help in the automation of the prediction of the breast cancer type and make the process more accurate.