A Performance of K-Nearest Neighbor Classification in Paraphilia Disease

Abstract

Paraphilia is still not widely known by the public. Lack of information about paraphilia is a serious concern of the Makassar City Government. This is because there are 12 types of paraphilia and some of them are contagious diseases such as fetishism, transvestism, sadomasochism, pedophilia, transsexualism, voyeurism, exhibitionism. There are several paraphilia diseases that are difficult to distinguish. The nature of paraphilia can be seen by society through its given (nature) and caused by environmental influences. In this study, the K-Nearest Neighbor (KNN) method has been applied to categorize the disease. The dataset used is derived from observations of 250 datasets. The dataset is divided into two, training data (165) and testing data (70). Based on the experiment, the k-NN method has an accuracy of Confusion Matrix of 8l%. On the other hand, the k-NN method is able to classify 12 venereal diseases quite accurately. Thus, this method was good as an alternative method for the classification task. For future research, optimization of the application will be performed to increase the accuracy of kNN.