K-Medoids Clustering untuk Pembentukan Database Stopword Bahasa Jawa

Abstract

Stopword is a word that can be ignored in the natural language process. This word removal process does not affect the text analysis process. The technique used to remove stopword is called Stopword Removal. This technique matches words to a stopword list. If the word is in the list it will be deleted. Javanese language to date still has a limited list of stopword. This study aims to form a list of stopword using cluster techniques namely K-medoids clustering. This technique groups words by occurrence in Javanese text. Each cluster result is tested by matching it with a stopword of javanese expert identification. The results of this study suggest that the stopword produced by k-medoids clustering with a value of K=13 has an accuracy of 70.5%.