An Advanced Approach to Forecasting Tourism Demand in Taiwan

Abstract

Forecasting has been considered important in a service industry. Many techniques have been applied to improve forecasting results. This study intends to apply a neural network based fuzzy time series model to forecast the international tourist numbers arriving in Taiwan. Neural network is good at handling nonlinear data. On the other hand, the fuzzy time series models have been applied to time series problems in various domains and have been shown to outperform some conventional models. The tourist numbers arriving in Taiwan show a nonlinear characteristic with a structural break. And the application of the neural network based fuzzy time series model is expected to outperform some other models in forecasting these tourist numbers.

Key Words:

• Fuzzy time series
• neural networks
• structure breaks
• tourist numbers