Breast Cancer and Prostate Cancer Detection using Classification Algorithms

Abstract

It is a known fact that, cancer rates have increased to great heights in the recent times. The only way to completely cure cancer is to detect its presence at an early stage, for which appropriate diagnosis is available. The main plot of this paper is the detection, and classification of such cancerous cells in the patient’s genome expression, on detection of which he/she can be provided rightful treatment. Modern day techniques have evolved that can help to detect the presence of cancer, such as Deep Learning, Artificial Neural Networks, Deep Convolution Networks, and Data Mining etc. In this paper we have dealt mainly with two types of cancer. Breast cancer and Prostate cancer in females and males respectively. We have implemented Machine Learning to find out signs of cancer, and what type of cancer, if seen. The reason being, physicians are capable of diagnosing a patient with cancer, with an accuracy of 71%, according to a latest research, on the other hand Machine Learning techniques can show up to 91% accuracy for rightful classification. Since, the primary focus is to detect and classify the type of cancer in the patient, we have used classification techniques/algorithms under Machine Learning such as Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR) and Naïve Bayes (NB). Effort has been made, to identify the best techniques providing the highest accuracy for both the cancers, and enhancing them with stratified K-fold and dimensionality