Abstract

Prostate cancer is a cancer that develops in the prostate in the male reproductive system, this occurs when prostate cells are attached to androgen receptors through the molecular docking process. The incidence of prostate cancer increases with age, where the risk that men have for prostate cancer in their lifetime is close to 10%. Early detection of cases of prostate cancer in many people or men who are susceptible to prostate cancer risk is important to start treatment and planning proper medical needs. One way that can be done in the detection of prostate cancer is to classify using a data mining approach with the Naïve Bayes algorithm and K-Nearest Neighbor (K-NN). This study aims to get the best classification results for detecting prostate cancer by comparing the two algorithms. The result of the accuracy of prostate cancer classification using naïve bayes algorithm is 80% and K-NN by 90%. Meanwhile, the overall precision value of Naïve Bayes and K-NN algorithms was at 71.5% and 93% respectively. The recall value for the Naïve bayes algorithm was 88% and the K-NN algorithm was 87.5%. Based on the accuracy, precision, and recall values of the two algorithms, the K-NN algorithm has a higher value compared to naïve Bayes' algorithm, so it can be said that the K-NN algorithm works well in classifying prostate cancer. Although Naïve Bayes' algorithm has a lower value compared to the K-NN algorithm, the average value for precision performance, recall, and accuracy is still above 70%. It can be said that Naïve Bayes' algorithm is quite good at classifying prostate cancer.