Aspect-Based Opinion Mining of Customer Reviews in the Hospitality Industry: Leveraging Recursive Neural Tensor Network Algorithm

Abstract

The hospitality industry is highly competitive, and in today's social media era, knowing customers' changing expectations has become essential. Accordingly, this study conducts aspect-based opinion mining (ABOM) on customer reviews to gain insights into these evolving perceptions and opinions. The study analyses user reviews from six prominent global chain hotels in South Asia, extracted from Tripadvisor. com. Employing Python for a robust text mining approach, we execute aspect-based opinion mining to extract nuanced sentiments expressed by users. Leveraging the Recursive Neural Tensor Network (RNTN) algorithm, widely used for sentence-level sentiment classification, we achieve a commendable average accuracy of 88.6% in categorising sentiments associated with specific words or aspects. This technical endeavour underscores the significance of aspect-based opinion mining and highlights the contributions of the proposed algorithm in illuminating insights into users' perceptions of hotel conditions. The obtained results stand to drive advancements in sentiment analysis techniques and provide a foundation for enhancing the quality of the hospitality industry in South Asia, thereby bolstering the region's tourism sector.