Hybrid modeling for forecasting domestic medical tourism demand in Tehran

Abstract

Introduction: One of the most important events in the tourism industry of each country is the demand for a product or destination and its true prediction of tourism. It should be noted that there are distances and deviations between actual values and predictions. The use of modern scientific and forecasting methods will make the results far more than an objective estimate and closer to the truth; this article pursues the same goal in the field of medical tourism. Methods: In the first step, factors affecting the demand for domestic medical tourism in Tehran were identified by 31 experts using Fuzzy Delphi and Dematel Fuzzy methods. The factors were then processed by MATLAB2017a software. After determining the demand function, and collecting monthly data of each effective factor from 2001 to 2015, three regression prediction models, a fuzzy neural network, and SVR algorithm were implemented using MATLAB software to measure and compare forecast errors. Results: The demand function for domestic medical tourism included: economic factors (individual income and wealth), service prices and cost of living in the destination, the cost of accommodation facilities, air pollution, and the price of alternative products (foreign travel), the number of medical centers, hospitals and laboratories. Conclusion: The proposed hybrid approach for regression and SVR algorithm can make a better prediction compared with the other methods of forecasting domestic medical tourism. Therefore, it is recommended to use the demand function and forecasting model to lower the forecast error while planning for domestic medical tourism demand in Tehran.