Penerapan Text Mining Menggunakan Algoritme Naïve Bayes Dalam Mengklasifikasi Sentimen Netizen di media sosial Twitter (Studi Kasus Pertemuan KTT G20 di Indonesia)

Abstract

The G20 is a forum that discusses global issues including Finance Track and Sherpa Track, this forum consists of developed and developing countries. The purpose of this research is to classify positive and negative tweets using the naive bayes method. The naive bayes method can predict future possibilities based on past experience and this method is recommended from several previous researchers because this method is considered suitable for analyzing positive and negative tweets in the dataset. The existing data is processed using rapidminer and produces a model. Based on the model of the naive bayes method seen from 700 training data, 675 positive predictions with class precision 100.00% and 22 negative predictions with class precision 88.00% and with an analysis accuracy value of 99.57%. So it can be concluded that many people have positive sentiments on twitter when the G20 is held in Indonesia and the G20 can be an effort to encourage the country's economy to be even better.