Classification of Hotel Reviews using Machine Learning Techniques

Abstract

The discussion and critique of products and services occur across various mediums, including the realm of social media. Reviews from previous customers offer a wealth of information about products, allowing shoppers to make pre-purchase product evaluations. Positive reviews are beneficial for a business because they can improve its reputation, attract new customers, and increase sales and profitability. Negative reviews impact businesses adversely but are just as important as a five-star rating. They provide opportunities to correct the issues to show potential clients that their opinions are cared for. Online reviews for hotels can be found on various websites. These opinions shared on social media platforms can shape the public's perception of the hotel. By using sentiment analysis and machine learning algorithms on existing hotel reviews, we create a classification system that can identify what the customer post on a hotel's website and thinks about the hotel. Logistic regression and Support Vector Machines (SVMs) were chosen as the classification algorithms. The classifiers are analyzed based on three criteria: accuracy, precision, and recall. The classifier with the best performance metrics is labelled the algorithm most suited for classifying hotel reviews.