A BACK-PROPAGATION NEURAL NETWORK WITH DELAY AND SHIFT WINDOW FOR TOURISM DEMAND FORECASTING

Abstract

This article studies machine learning techniques and factors that affect tourism demand to develop a predictive model for tourism demand in the coming years. The model was developed using the back-propagation neural network approach and expert knowledge for analyzing factors affecting tourist satisfaction. The data used in the study were collected over a ten-year period and comprised information on the local economic and social situation, as well as specialized tourism data. In addition, survey results evaluating tourism in An Giang province in 2019 were included. The study results demonstrate that the developed model has successfully captured the underlying patterns in the An Giang tourism data, enabling the prediction of the necessary tourism indicators for the future. The model achieved a high level of accuracy with an RSME of 0.04. Furthermore, our approach showed several advantages when compared to other classical statistical methods. Based on our research findings, we proposed policies to support businesses, planning, and management units in forecasting and investing in the development of tourism in each specific locality more effectively.