Sentiment Analysis towards Hotel Reviews

Abstract

The used of social media analytic towards the information in social media could bring a significant impact toward the hotel business thus making it easier to gain such insight and analysis. A preliminary study showed that the hoteliers in Malaysia are unaware on the sentiment analysis computation in analyzing the customer review through TripAdvisor despite having the website to promote their hotel. This study aims to provide another addition of sentiment analysis study especially on hotel reviews dataset gain from TripAdvisor. The process start by scrapping the reviews from The Majestic Hotel in TripAdvisor. The dataset then undergoes a pre-processing stage using Valence Aware Dictionary for Sentiment Reasoning to identify the sentiment in the text reviews. The text reviews are converted into vector form and fed into machine learning classifier. Three type of machine learning classifier was chosen to classify the review that are Naïve Bayes, Support Vector Machine and Maximum Entropy. The classifiers are then evaluated using three measurement from the accuracy, precision and recall value perspectives in order to find the best classifier. Based on this study, Maximum Entropy gained the highest value compared to the other two classifier.