Sentiment Analysis Of Hotel Customer Reviews Using K-Nearest Neighbor (K-NN) Method (Case Study: Hotels.com, Booking.com, Agoda.com)

Safitri, Rahma Nimas
Nurcahyawati, Vivine
Lemantara, Julianto

Jurnal Sistem Informasi Universitas Dinamika
Vol 9, No 1 (2020)

With the increasing number of gadgets and other online media, it is possible for consumers to provide reviews of services in the form of comments and opinions. In Appgrooves only assess by rating, but sometimes the rating is not enough to show consumers' responses to the services they get. From these problems, sentiment analysis is needed to classify user reviews based on positive and negative sentiments. In this study using the K-Nearest Neighbor method to classify negative and positive reviews. The reason for using the K-Nearest Neighbor algorithm is because the level of accuracy is good and effective when used on training data which is large and contains information that is less or not meaningful (noisy). Based on the validity test using 10-Fold Cross Validation, the accuracy for Hotels.com is 94.55%, for Booking.com is 87.58%, and for Agoda.com is 98.83%.