Aspect-based Opinion Mining for Code-Mixed Restaurant Reviews in Indonesia

Abstract

The goal of opinion mining is to extract the sentiment, emotions, or judgement of reviews and classified it. These reviews are very important because they can affect the decision-making from a person. In this paper, we conducted an aspect-based opinion mining research using customer reviews of restaurants in Indonesia and we focused into analyzing the code-mixed dataset. The evaluation conducted by making four scenarios namely removing stopwords without stemming, without removing stopwords but with stemming, without removing stopwords and stemming, and preprocessing with removing stopwords and stemming. We compared five algorithms which are Random Forest (RF), Multinomial Naive Bayes (NB), Logistic Regression (LR), Decision Tree (DT), and Extra Tree classifier (ET). The models were evaluated by using 10 folds cross validation, and the results show that all aspects achieved highest scores with different algorithms. LR achieved highest score for food (81.76%) and ambience (77.29%) aspects while the highest score for price (78.71%) and service (85.07%) aspects were obtained by DT.