Comparative analysis of RFE in Naïve Bayes classification in Breast Cancer Detection

Breast Cancer is a major reason for increasing mortality rate among women. There has been exponential growth in the number of women suffering from the disease. An early and precise detection is the necessary as to prevent breast cancer successfully. Data mining techniques have an astounding potential to develop a system which can be utilitarian in cancer detection. To classify benign and malignant tumor we have used Naïve Bayes classification technique and further improved its accuracy by applying Recursive Feature Elimination. This paper is a comparative study on the implementation of Naïve Bayes Classifier and further improving its accuracy by applying Recursive Feature elimination on the same classifier, The classifier was implemented on the Wisconsin Diagnosis Breast Cancer (WDBC) dataset. Our experiments have shown that when Naïve Bayes was used for predictive analysis, it had an accuracy of 94.20% and an increased accuracy of 95.95% after applying Recursive feature elimination method.