Prostate Cancer Detection Using Different Classification Techniques

摘要

Prostate cancer is a widespread disease among the male population. Its early diagnosis and prognosis are challenging tasks for clinical researchers due to the lack of very precise, fast and human error free diagnostic method. The purpose of this research is to develop a novel prototype of clinical management in diagnosis and management of patients with prostate cancer. Various classification algorithms were applied on a cancer database to devise methods that can best predict the cancer occurrence. However, the accuracy of such methods differs depending on the classification algorithm used. Identifying the best classification algorithm among those available is a difficult task. In this paper, the results of a comprehensive comparative analysis of nine different classification algorithms are presented and their performance evaluated. The results indicate that none of the classifiers outperformed all others in terms of accuracy, meaning that multiple classifiers can serve clinicians in diagnostic procedure.