Study On Domestic Tourism Demand Forecasting Based On Fuzzy Theory

In recent years, with the sustained and rapid development of tourism industry, tourism has become one of the industrial which has the most powerful development momentum in the world. The strong development of tourism industry led to the rapid development of economy, more and more countries develop the tourism industry and put it as a pillar industry, and hope it can promote the development of the social economy. The tourism demand forecasting can provide a reference for national tourism administration to make the strategic planning and policy, also can provide a reference for the tourism enterprises to reform and develop, so the tourism demand forecasting can guide the optimal allocation of market resources.The particularity of tourism product determines the tourism demand is easy to be affected by many factors, so travel demand’s prediction becomes more complex, there has no a better demand forecasting method to treat the complex influencing factors. This paper aims to explore a more suitable forecasting method for tourist complex environment to improve the accuracy of predictions, find the law of development of domestic tourism and thereby predict future trends, better serve domestic tourism management and tourism decision-making etc.This paper improves the fuzzy time series model in forecasting tourism demand with the same interval method for domain partition problem, raises fuzzy clustering algorithm for the domain’s non equal division. Then, in view of the defect that the traditional theory of grey prediction model is more easily affected by the changes of the object of study sample data, combined with the advantages of markov chain method, merged into fuzzy classification theory and used it in the latter part of forecasting, the fuzzy grey markov chain method forecasting model is proposed. Finally, most of the current domestic tourism demand forecasting method using a single forecast, with low prediction accuracy and stability, introduces a kind of induced ordered weighted averaging operator method, uses the separate prediction model established by Chapters II and III, establishes a comprehensive predictive model based on IOWA operator. The research results show that, the improved fuzzy time series model not only ensure the high precision of prediction and simplifies the calculation, but also avoid the error caused by setting clustering number. The fuzzy Grey Markov chain prediction model can reflect the trend of development of historical data, and ensure the high precision of prediction when historical data are fluctuation. The combination model covers more information, overcomes the defects of single model prediction, and considers the change of single model’s prediction accuracy in different period, the prediction accuracy and stability has been further improved.