Machine Learning Techniques for Sentiment Analysis of Hotel Reviews

Abstract

In today's scenario online reviews on various digital platforms plays a vital role for customers to buy products. Based on the reviews and ratings by the consumer on E-commerce platform like flipkart, amazon etc. products are widely accepted or rejected. Apart from products people also look for the reviews of the services provided from restaurants, hotels, airlines etc. Sentiment analysis helps the developers to easily analyze the reviews and categorize them as positive or negative. In this paper, service of a hotel is analyzed by finding out the polarity of the reviews in order to get the subject information. Aspect detection and sentiment classification are the main tasks focused here. For aspect detection latent dirichlet allocation (LDA) is used for building the topics. Different machine learning classifiers like naive bayes classifier, SVM, decision tree and logistic regression are used for classification of reviews. Evaluation is done by computing the accuracy, recall, precision and F score of these algorithms.