A comparitive analysis on various machine learning techniques for breast cancer detection

The primary goal of the study is to detect breast cancer patients at an early stage. It is the mass elegant deadly diseases compared to rest of the infections in medical field. Breast cancer was one of the reasons for dying among women globally on report of the WORLD HEALTH ORGANIZATION (WHO). This research seeks to classify tumors based on data attributes as malignant (cancerous) tumors that penetrate and destroy healthy body tissues or benign tumors that do not (non-cancerous). We are using a support vector machine (SVM) classifier, random forest (RF) classifier & decision tree classifier for breast cancer detection by training, testing, and splitting its attributes. For detecting the breast cancer we are using the Wisconsin Breast Cancer Dataset (WBCD) data obtained out from UCI library. The database composed of 569 analyses, out of which 357 are benign (non-cancerous), 212 are malignant (cancerous). The comparative analysis of support vector machines (SVM) is done for finding the accuracy. The method gives good results for the data of training. Moreover, the approach attains accuracy 96.50% by Support Vector Machine, 97.12% by Random Forest Classifier and 99.50% by Decision tree Classifier.