Recommendation system for customer preferred mental healthcare facility

Mental health is a specialty of healthcare segment which refers to the psychological wellbeing of a person. A key issue faced by patients with mental health illnesses, is the difficulty of finding a suitable hospital subject to various given conditions such as demographic characteristics, type of treatments, quality of services, and affordability. Most of the well-known hospital ranking systems in USA have considered a limited range of psychiatric hospitals and such rankings have been biased towards hospital reputation with absence of patients' needs. As a matter of fact, it does not necessarily provide customer requirements related information. Data mining can play a significant role in mental healthcare by benefiting both facilities and patients. Therefore, the objective of this research is to develop a predictive model to recommend facilities which are well-suited for actual patients' needs. Data from National Mental Health Services Survey is employed for this research. The healthcare facilities are initially clustered based on the fulfillment of customer oriented dimensions. Then, the supervised learning techniques are applied to predict the ideal cluster of healthcare facilities for a given set of patients' needs. We provide the performance evaluation metrics to demonstrate our proposed models' applicability for addressing the current issues.