Analisis Sentimen Pada Review Produk Kosmetik Bahasa Indonesia Dengan Metode Naive Bayes

Abstract

Sentiment analysis is a computational study of the opinions, behaviors and emotions of people toward the entity which in this research is a product review. By reading the review of products based on the experiences of other consumers, it will be recognized the quality of a product. As cosmetic products on the market are very diverse, both in terms of type and brand. However, not all cosmetics have good quality and it is to be noticed by the consumer. So, reexamination of the cosmetic product review by classifying these reviews into positive and negative class is an excellent way to determine the response of other consumers about the product quickly and accurately. The classification technique that mostly used is Naive Bayes. Naive Bayes is a popular classification technique for text because very simple, efficient, and has good performance in many subject. But, Naive Bayes has weaknesses that is too sensitive when there are too many features which cause wrong classfication. Because of that, in this research tf-idf is used as feature selection to improve Naive Bayes performance. This research aim is to produce a system that can classify reviews to their respective positive or negative class. The accuracy measurement of Naive Bayes is done by using confusion matrix and indicate the results of average accuracy from 69% to 82%.