Comparision of Four Machine Learning Techniques for the Prediction of Prostate Cancer Survivability

Abstract

Prostate cancer is regarded as the most prevalent cancer in the word and the main cause of deaths worldwide. Many traditional machine learning classification techniques has been applied to prostate patient survivability prediction, such as k Nearest Neighbors (KNN), Decision Tree (DT), Naïve Bayes (NB)and Support Vector Machine (SVM). In recent years, deep learning has been proved as a strong technique and became a research hotspot. As a kind of deep learning method, in this study, artificial neural network and several traditional machine learning techniques are applied to SEER (the Surveillance, Epidemiology, and End Result program)database to classify mortality rate in two categories including less than 60 months and more than 60 months. The result shows that neural network has the best accuracy (85.64%)in predicting survivability of prostate cancer patients.