Abstract
Medical tourism involves individuals traveling internationally to receive medical treatments, surgeries, or wellness therapies, often due to lower costs, shorter wait times, or access to specialized care. Yet, occupancy of rooms in medical tourism is plagued with inefficiencies like overbooking during peak periods and underutilization in off-peak periods. The inefficiencies affect patient satisfaction and hospital income. An Automated Meta-Machine Learning (AMML) model was proposed for predicting room occupancy and patient satisfaction for medical tourism. A hedonic value perception framework was applied to measure tourists’ behavioural intentions, derived from 150 participants. With this behavioural information, the AMML model utilizes automated machine learning and the ARIMA model to make timeseries forecasts. A regression classifier also smooth forecasts on room occupancy and satisfaction metrics. A mean prediction error of a mere 0.8 days in occupancy forecasting and accuracy levels of 92% for total satisfaction, 89% for quality treatment, and 85% for facility cleanliness were realized. Although limited inconsistencies were observed in cost satisfaction and staff professionalism, good performance was exhibited in forecasting recommendation probability, achieving 90% accuracy. Hospital planning, resource optimization, and improvement of patient satisfaction in the medical tourism business are improved by this strategy.
Keywords: Automated Classification, Behaviour Intention, Machine Learning, Medical Tourism, Prediction, Room Occupancy