The Study Of Tourism Demand Forecasting Based On The Grey-neural Network Theory

Since the Chinese economic reform, as a high-yield, low-input and one of theindustry’s global economic which rapidest development—tourism is mushrooming upthroughout the country. The sustainable development is mainly due to the comprehensivenational strength has greatly enhanced, dweller income rise steadily and related supportpolicies, but also inseparable from the tourism department development the tourismeconomy, including tourism destination organization ability, convenient transportation andtourism facilities and gradually improve the ability. As a high output, low input sunriseindustry, tourism plays an essential role in the development of tourism resources utilizationand construct, expand inside need, national economy sustainable and rapidly development.Therefore, the deep research of domestic tourism needs, the accurate analysis of the currentand predict future changes of domestic tourism market, will become the arduous task in thetourism management, plan and development.Up to now, the tourism demand forecasting model which has been established mainlyinclude the sequence model and the regressive model, however, the gray models andartificial neural network model which widely used in other areas haven’t been appliedwidely of is applied. Although these forecasting models can be used in this area, but whichkind of also has no coherent file format. In this paper, the gray system theory, the Markovmethod, neural network theory and the combination of modeling methods are used tomodel and analyze the demand for domestic tourism demand.First, this paper used the qualitative analysis methods analysis the four factors oftourism demand, and reveal the mechanism of the number of domestic tourists andresidents’ income, the relationship between tourism services, tourism, environment androad traffic conditions.Second, the grey correlation analysis made quantitative description of each factor onthe size of the number of domestic tourists. And used the grey GM(1,1) model andGM-Markov model to make tourism demand time series forecasting.Again from the travel demand time sequence and influence factors of multiplesequence start with BP neural network to predict.Finally, make the BP neural network and GM-Markov modified model organiccombination, establish the combined gray neural network model for prediction of thenumber of tourists and tourism the number of actual time series data to analyze the precision of the model. With the single BP neural network and gray GM (1,1) model andother prediction methods compared. Use MAPE as the accuracy indicators of predictmodels, the results showed that: GM-Markov is2.5134, the single BPNN is1.5224,multivariable BPNN is1.9509, PGNN is1.4085, SGNN is1.5993.Thus reflecting the combined model can take the public long to make up self ownshort, and has less sequence information of the time required for the dual advantages ofhigh prediction accuracy, its predictions with a certain reference value.