Cluster Analysis of Indonesian Hospital Service Quality Using Online Review Mining

Hospital service quality has become a primary focus in the global healthcare industry. While technical quality remained crucial, patients often assess service quality based on interpersonal interaction and surrounding environment. This research aims to consolidate hospital data into a comprehensive dataset, enabling the utilization of machine learning techniques to cluster hospitals and identify factors differentiating service quality automatically. This research analyzes patient reviews of hospitals in Surabaya, Indonesia, by sentiment analysis, Latent Dirichlet Allocation, and K-means clustering. The result of this research classifies four clusters based on tangible, reliability, responsiveness, assurance, empathy, sentiment score, and rating. Reliability and empathy are marginally significant in forming four clusters. The refined clustering finds three clusters due to the existence of clusters two and four with similar values. Consequently, this implies that hospital classifications do not fully capture the level of service quality provided, underscoring the necessity for a deeper exploration based on patient experiences.