Forecasting Foreign Tourists in Northern Thailand Using Artificial Neural Network

Abstract

Tourism industry is a major economic factor of Thailand, make substantial revenues country, employment of the citizens in urban and local, has developed country,
investment in transportation infrastructure and other. Accurate forecasting of tourism demand is important in tourism planning by both the nation policy development and business sectors owing to the limited resources of employment and investment. This paper
presents a model to forecast foreign tourists visiting Thailand in the northern using Back
Propagation Neural Network (BPNN) compare with Recurrent Neural Network (RNN). For the
benefit of tourism products and services e.g. airlines seats, transportation, hotels room, tourist attractions, travel companies and food. We have studied the comparison of the number of neurons in the input layer 2 to 6 nodes according to the time lag and the
hidden layer 2 to 10 nodes. The performance of forecasting model evaluation employs the
mean absolute percentage error (MAPE). The results show that the model of neural network, which is high performance, the BPNN is a model 6-9-1. It has 6 input and 9 hidden odes, has a MAPE of 22.16%, the RNN is a model 6-9-1. It has 6 input and 9 hidden nodes,
has a MAPE of 11.02%.