Overall and Disease-Free Survival Prediction of Postoperative Breast Cancer Patients using Machine Learning Techniques

Cancer is a disease which is caused by continual unrestrained proliferation of tissues. Breast cancer is a fatal disease commonly found in females. Breast cancer cure is possible with timely diagnosis, prediction of survivability and its treatment. Prediction of survival from breast cancer on basis of patient historical record is a challenging task. The primary goal of this research is to predict the overall survival (OS) and number of years a patient can survive without any symptoms of breast cancer called disease free survival (DFS). For the experimentation, data set of female patients of Pakistan diagnosed with breast cancer have been considered. Six (6) machine learning classifiers have been trained on the data gathered from INMOL hospital of Pakistan. The classifiers include Naïve Bayes, J48 decision tree, SVM, Random forest, AdaBoost and JRip. To the best of our knowledge no such work has been done on the Pakistani female breast cancer data. From the results of experimentation, it has been found that the performance of JRip for both OS and DFS is much better than the others. This prediction will be helpful for patients and paramedical staff (doctors) to foresee and predict the situation which may befall.