Comparative Study on Machine Learning Algorithms for Breast Cancer Diagnosis

Breast Cancer (BC) is the second most found cancer in women worldwide. Different symptoms to identify breast tumors are pain in the breast, lumps, and a change in skin color and texture. In this study, we used mammograms for Breast cancer classification. The radiologist uses computer-aided diagnosis tools for a second opinion diagnosis. Nowadays, Machine learning algorithms are used for the advancement of technology. The proposed research mainly focuses on Machine learning classification algorithms for classifying data as benign and malignant. The datasets used are MIAS and Wisconsin for Breast cancer classification. The algorithms used are Logistic Regression, Naive Bays classifier, Support Vector Machine, Decision tree, and Random Forest. Augmentation was applied to increase the size of the MIAS dataset. The performance of different classifiers is compared using measures like accuracy, precision, recall, and F-measure. Results show that for the MIAS dataset, Random forest gives the highest accuracy of 0.73 compared to all other algorithms. For the Wisconsin dataset, Logistic regression out-performs with 96.49 per cent accuracy in classifying the data.