Open AccessNEWS TOPIC CLASSIFICATION ON TRIBUNNEWS ONLINE MEDIA USING K-NEAREST NEIGHBOR ALGORITHM
Author(s) -
nfn Herman
Publication year - 2018
Publication title -
journal of information technology and its utilization
Language(s) - English
Resource type - Journals
ISSN - 2654-802X
DOI - 10.30818/jitu.1.2.1879
Subject(s) - computer science , k nearest neighbors algorithm , information retrieval , test (biology) , algorithm , data mining , artificial intelligence , biology , paleontology
Online media journalists like tribunnews journalists usually determine the news category when make news input. Unfortunately, often the topic submitted is not in accordance with what is expected by the editor. These errors will make it difficult for news searches by customers. To eliminate these errors, editors can be assisted by an application that able to classify topics. Thus, editors is no longer too dependent on journalist input. This study aims to design applications that able to classify topics based on the texts contained in the news. The method used is the K-Nearest Neighboor algorithm. This design has produced a system that able to classify news topics automatically. To measure the accuracy of the application, several test were carried out by comparing between its results and the results of manual classification by the editor. The tests those carried out with several scenarios produce an accuracy rate of 82%