The prediction of prostate specific antigen (PSA) using different models of artificial neural networks

In this study, prostate specific antigen (PSA) values in patients with prostate cancer and without prostate cancer were predicted with the help of differenet artificial neural networks (ANN) models. In the first stage of this study, information on statistics and artificial neural networks were presented about the concept and history. Then, the basic structure and working styles of some artificial neural network models were investigated. Literature search was carried out for various areas of artificial neural networks? applications.In the application stage, the data regarding 203 men including age, DRE, TPSA, fPSA, f/t, Tvolüm, TZvolüm, PSAD, TZPSAD and pathology were provided retrospectively in the outpatient clinic, Department of Urology, Faculty of Medicine, Gazi University. The models of artificial neural networks were used in predicting PSA values. ANN models were trained with different training algorithms and related predictions were made using these different ANN models.