Implementing Rule-based and Naive Bayes Algorithm on Incremental Sentiment Analysis System for Indonesian Online Transportation Services Review

Abstract

The emerging trend on smartphone application and service use on a daily basis, has also increased the volume of online opinion regarding various topics on the internet. In Indonesia, one of the most popular topics to share, post and comment is online-based transportation service (TNCs). These comments could lead to valuable knowledge that would be tremendous assets for supporting critical business intelligence applications. The knowledge gained from social media can potentially lead to the development of novel services that are better tailored to users’ needs and also meet the objectives of businesses offering them. The problem to build an effective Indonesian sentiment analysis system is that there is still no availability of the corpus, complete with each word characteristic, whether it subjective, adjective, adverb, noun, etc. Another problem is because their cultural heritage, or for politeness reason, Indonesian people often used negation in their sentence. So instead of saying “ugly”, they say “not good”, or instead of saying expensive, they said “not cheap”, which could lead to miss-classify of the sentiment. Thus, this research focus on building model that has the ability to classify TNC element target on its sentiment class, by considering negation form sentences and then implement it in the proposed sentiment analysis system. Another important feature is system’s ability to learning new keywords for TNC elements and sentiment. This proposed approach would use rule based algorithm to classify target object, the polarity of sentiment and negation from online opinion. And used Naive Bayes algorithm for the incremental feature. Result from this study show that the proposed system is able to classify user opinion with 90% precision and 70% recall. This concludes that from evaluation results, the proposed algorithm performs well to automatically analyze sentiment.