Using Machine Learning Techniques to Maximize Profitability in the Hospitality Industry

In the highly competitive hospitality industry, customer satisfaction is critical, and online reviews and ratings are instrumental in influencing hotel booking decisions. Tourists prioritize hotels offering exceptional services, making it vital for hospitality businesses to understand and address customer concerns promptly and effectively. Machine learning algorithms have become essential tools for analyzing online reviews and ratings. This paper analyzes over 70,000 reviews from TripAdvisor, focusing particularly on negative reviews with low ratings. The proposed methodology encompasses both unsupervised and supervised learning Methods. The unsupervised learning phase utilizes the K-Means algorithm to cluster similar reviews and identify prevalent themes and issues. The supervised learning phase employs five distinct algorithms: Support Vector Machine (SVM), Naive Bayes, XGBoost, Logistic Regression, and Random Forest, to classify reviews. These algorithms are evaluated on accuracy, precision, recall, and F-score metrics. Among them, the SVM algorithm demonstrates superior performance, achieving an accuracy rate of 76%%. The findings reveal that SVM is highly effective in analyzing online reviews, offering valuable insights for strategic decision-making in the hospitality sector. The results suggest practical implications for hospitality businesses in enhancing service quality and customer satisfaction.