Paper Title
A Machine Learning-Based Holistic Approach to Predict the Survival of Breast Cancer Patients

In the early stages of breast cancer, inefficient treatment methods, as well as the patient's health condition may impact the patient's lifetime expectancy. In this study, given a set of explanatory variables that include the patient's demographics, health condition, and cancer treatment regimen, our objective is to investigate the performance of four different machine learning methods including an artificial neural network, Support Vector Machines and Random Forests. To achieve this ultimate goal, we utilize these three methods with a ten-fold cross validation to predict the one year, five years, ten years and fifteen years survivability of the breast cancer patients after initial diagnosis. The results of each method are compared with respect to accuracy, sensitivity, specificity, and area-under-the-curve (AUC) metrics. According to the proposed methodology, we observe that the Random Forest method shows better performance when compared to the others in most of the evaluation criteria that have been used in this study. In addition, in all prediction models, the stage of the cancer has been determined as the most important predictor of breast cancer survivability. The current study can be utilized by the medical practitioners as well as medical researchers for potential prospective studies. Needless to say, such outcomes can be considered as a decision support mechanism, not as a primary decision maker. Index Terms – Artificial Intelligence, Breast Cancer, Data Mining, Machine Learning, Prediction