ASSESSING SERVICE QUALITY BASED ON CUSTOMER REVIEWS USING MACHINE LEARNING TECHNIQUES: A CASE STUDY OF THE HOSPITALITY SECTOR

Abstract

This study assesses service quality in the hospitality sector by mining and analysing customer hotel reviews with machine learning techniques. It aims to identify the key service‐quality dimensions that drive guest satisfaction and to provide hotel managers with actionable, data‐driven insights. We compiled a dataset of 28,945 English‐language hotel reviews. Natural language processing and sentiment‐analysis methods were applied to extract textual features and polarity scores. Topic modelling was used to surface latent themes, and two supervised classification models—support vector machines (SVM) and random forests—were trained to predict sentiment labels and highlight critical service dimensions. The analysis revealed four principal dimensions of service quality: cleanliness, staff behaviour, amenities, and value for money. Both SVM and random‐forest classifiers achieved robust predictive performance (accuracy > 85 %), validating the suitability of a multi‐method framework. Unlike prior work that focuses narrowly on sentiment analysis or small samples, this study leverages a large review corpus and integrates sentiment analysis, topic modelling, and supervised learning into a unified framework. It offers a scalable approach and discusses potential extensions—such as deep‐learning architectures, multilingual analysis, and real‐time monitoring systems—to further advance data‐driven service‐quality assessment