Model for the Classification of Feelings of Reviews Using Techniques Decision Trees Case Studies Hotel Reservation Web Site

Abstract

Sentiment analysis with text data mining is widely used for business success. The researcher’s study is to identify the customer review sentiment from Booking.com and Traveloka.com websites that both websites have indicated the customer reviews and we cross-check in the prepare data process. We used more tokenize engine including included MM, NEWMM andLONGEST with those of cut word suite pythainlp’s stopword and our custom stopword. The results show that the NEWMM engine has accuracy at 0.90 and 0.90for precision both positive and negative reviews. However, the NEWMM engine with pythainlp’s stopword has accuracy at only 0.81 and precision at only 0.81. Concluding, our sentiment model with the decision tree technique can found better performance with the NEWMM engine and research’s custom stopword.