Klasifikasi Abstrak Jurnal Repositor di Teknik Informatika UMM Menggunakan Metode Neighbor Weighted K-Nearest Neighbor

Abstract

Journals or scientific articles are text data in the form of written works that have been made by researchers who conduct research on a topic in a particular field. In general this scientific journal is published online by scientific information provider sites and universities. In this research, take a case study UMM Repositor Informatics Engineering journal which is still not categorized based on the fields of interest such as RPL, Network, Game, and Data Science. This study uses the Neighbor Weighted K-Nearest Neighbor (NWKNN) method, a modification of the KNN method in general. The classification results using this method show that the optimal k value is at number 17 with e = 2 which has 60,25641026% of accuracy. Due to that according to previous studies using the NWKNN method, if k is of great value then a test data will have a big probability of true entering the class category with the most highest score calculate.