Evaluation of Prostate Cancer via Machine Learning

Abstract

By training computers with machine learning technique, patients can be prevented from being exposed to unnecessarily difficult examinations. In recent years, machine learning-based disease assessment approach has gained importance in terms of the benefits it provides to clinical methods. There is a remarkable increase in studies in this direction. There are a limited number of clinical guiding parameters in predicting some types of cancer, and this limitation pushes the patients under treatment to a very frustrating process. For this reason, apart from ordinary procedure of the traditional medicine, an alternative approach to predict the any type of cancer is making a computer-based evaluation that has become a highly studied method in recent years. In this study, a machine learning (ML) approach will be used to evaluate prostate cancer, which is the second most common cancer-related death in men worldwide. For this purpose, the K-Nearest Neighbor (kNN) algorithm based on ML will be used with feature selection, which is a dimension reduction technique. An open source database, Kaggle, was used for the evaluation. The accuracy value of the used algorithm was found 88%.