Analisis Sentimen Opini Masyarakat terhadap Institusi Polri pada Media Sosial Twitter menggunakan Metode Support Vector Machine dan Naive Bayes

Abstract

Sentiment analysis is the process of extracting text data to obtain information about positive and negative sentiments. This method helps to study sentiments from various kinds of social media content in the form of tweets about the institution of the Republic of Indonesia's state police.Support Vector Machine is a supervised learning technique, has a good level of accuracy and quality. However, its implementation requires a sequential training stage and must go through a testing process. The advantages of the Support Vector Machine method can identify separate hyperplanes so that they can maximize the margins of different classes. The drawback of this method is that problems that have the same features can significantly affect the level of accuracy. From a dataset of 1100 obtained results of accuracy between 85% to 86% with the results of the accuracy values of the four testing trials of 90%: 10%, 80%: 20%, 70%: 30%, and 60%: 40%. Naïve Bayes is a clustering algorithm, whose origins are from Bayes' theorem, whose process consists of computing probability values for text data, and which can process large amounts of data with a high degree of accuracy. From a dataset of 1100 obtained results of accuracy between 79% to 85% with the results of the accuracy values of the four testing trials of 90%: 10%, 80%: 20%, 70%: 30%, and 60%: 40%.