DIAGNOSIS OF BREAST CANCER USING MACHINE LEARNING ALGORITHMS

Abstract

 Breast cancer is the main cause of the death in women. This study aims at diagnosing breast cancer at early stages with machine learning algorithms. Early diagnosis of breast cancer helps in preventing death. This problem is addressed by using  machine learning algorithms which can accurately classify the breast tumor as Malignant or Benign. Breast cancer diagnosis is done with the help of Wisconsin Breast Cancer dataset. This dataset is widely used for classification process in diagnosing breast cancer. This dataset is classified as 65% training data and 35% testing data and machine learning algorithms is used for classifying breast cancer tumors. A comparative study on various classification approaches such as Random Forest classifier, Logistic Regression, Naïve Bayes is done with a focus on finding the best approach. It has been observed that the selection of parameters play a very important role in classification process. The study shows that Random Forest classifier gives the best results. On these datasets we obtained accuracy of 96.5% on average. This is compromising when compared to previous observations.