Analisis Sentimen Berbasis Aspek pada Wisata Halal dengan Metode Deep Learning

Abstract

Halal tourism is becoming a trend today, along with the improvement of the number of Muslim tourists. Halal tourism is a part of the tourism sector, which offers service by referring to Islam rules. By halal tourism development, start to occur tourism place-review related to the facility that can facilitate Muslim tourists. The facilities involve toilet cleanliness, worship place availability, and halal food availability. However, unfortunately, there are very few who analyze the review regarding this issue. Therefore, through this research, it was done a sentiment analysis based on various aspects of tourism places in Asian countries using the deep learning method. This method was used because it could generate a good performance accuracy. The data used were English reviews taken from the TripAdvisor website. The data then worked and processed so that it could recognize the sentiment and aspect from the review as well. There were three aspects used those were a mosque, halal food, and toilet. After a test was done, the CNN method obtained the highest accuracy result if it was compared with the other methods, both from aspect classification or sentiment classification. With the CNN method, the aspect classification produced an accuracy of 98.299%. Meanwhile, the sentiment classification gained an accuracy of 93.96%. The result of this research is expected to help develop a strategy to promote halal food more.