The sentiment analysis of Indonesia commuter line using machine learning based on twitter data

Abstract

This paper presents The Sentiment Analysis of Indonesia Commuter Line (KRL) using Twitter data. Some of people are expressing their complaints about public transportation, especially the satisfaction of Commuter Line service on Twitter. We collected the data from Twitter to classify many public opinions into positive or negative label then Machine Learning model will be used as classification model to classify positive or negative opinion. Multinomial Naive Bayes (MNB), Random Forest (RF) and Support Vector Machine (SVM) is used as a model, then we measure all of model performances. And the result is Support Vector Machine produced has the highest accuracy of all with 85%. Net Sentiment Score (NSS) is also computed in order to determine whether KRL features meet customer's satisfaction.