基于灰色系统和神经网络的旅游需求预测

发布时间：2018-09-12 14:07

【摘要】：改革开放后，旅游业作为一项高产出、低投入、全球经济发展最快的朝阳产业，如雨后春笋般在全国各地悄然崛起。伴随着社会经济的稳步发展，旅游需求也在不断增长。深入研究国内旅游需求、准确分析当前需求现状并预测未来国内旅游市场的变化，，不仅对旅游资源的开发利用和环境建设相当重要，对于政府扩大内需和相关扶持性政策的制定以及国民经济持续快速发展也至关重要。目前已建立的惯用的旅游需求模型包括时间序列预测模型和回归模型，对于其他领域广泛应用的灰色模型和人工神经网络模型则应用不多。尽管这些模型均可用于这方面预测，但具体运用哪一种进行旅游预测分析始终没有一个固定的、统一的套用格式。本文采用灰色系统理论、马尔科夫方法、BP神经网络理论和组合模型方法对国内旅游需求进行建模和分析。首先采用定性分析法对旅游需求的四个影响因素做了分析，从机理上揭示出国内旅游人数与居民收入、旅游服务、旅游环境和道路交通状况之间的关系。其次采用灰关联分析法定量地描述各影响因素对国内旅游人数的影响程度大小。并用GM(1,1)模型和GM-Markov模型对旅游需求的时间序列进行预测。再次从旅游需求时间序列本身和影响因素多序列各自出发用BP神经网络模型进行预测。最后将BP神经网络与用马尔科夫修正的灰色模型进行有机地结合，建立了灰色神经网络组合模型，用于旅游人数预测，通过旅游人数的实际时间序列资料来分析模型的预测精度。并与单一的BP神经网络和灰色GM(1,1)模型等预测方法进行比较。结果表明：GM-Markov模型预测的MAPE为2.5134，单变量BPNN模型预测的MAPE为1.5224，多变量BPNN模型预测的MAPE为1.9509，三种模型并联后的组合模型预测的MAPE为1.4085，串联后的组合模型预测的MAPE为1.5993。由此反映出组合模型能够取众长补己短，且具有所需时间序列信息量少与预测精度高等双重优点，其预测结果具有一定的参考价值。
[Abstract]:After the reform and opening up, tourism as a high output, low investment, the world's fastest economic development sunrise industry, such as bamboo shoots in all over the country quietly rising. With the steady development of social economy, tourism demand is also increasing. It is very important not only for the development and utilization of tourism resources and the construction of environment, but also for the further study of domestic tourism demand, the accurate analysis of the current situation of demand and the prediction of the changes in the domestic tourism market in the future. It is also important for the government to expand domestic demand and related supportive policies, as well as the sustained and rapid development of the national economy. The established tourism demand models include time series prediction model and regression model, but they are not widely used in other fields such as grey model and artificial neural network model. Although these models can be used to predict this aspect, there is no fixed and uniform format for tourism prediction analysis. In this paper, the grey system theory, Markov method, BP neural network theory and combination model method are used to model and analyze the domestic tourism demand. Firstly, the qualitative analysis method is used to analyze the four influencing factors of tourism demand, and the relationship between domestic tourism population and residents' income, tourism service, tourism environment and road traffic condition is revealed from the mechanism. Secondly, grey correlation analysis is used to quantitatively describe the influence of various factors on the number of domestic tourism. GM model and GM-Markov model are used to predict the time series of tourism demand. Thirdly, the BP neural network model is used to predict the time series of tourism demand and the multiple series of influencing factors. Finally, combining BP neural network with Markov modified grey model organically, a grey neural network combination model is established, which can be used to predict the number of tourists. The prediction accuracy of the model is analyzed by using the actual time series data of the tourist population. It is compared with single BP neural network and grey GM (1 ~ 1) model. The results show that the MAPE predicted by the w / GM-Markov model is 2.5134, the MAPE predicted by the single variable BPNN model is 1.5224, the MAPE predicted by the multivariable BPNN model is 1.9509, the MAPE predicted by the combined model is 1.4085, and the MAPE predicted by the series combined model is 1.5993. The result shows that the combined model can compensate for the short and the short, and has the advantages of less time series information and high prediction accuracy. The prediction results have certain reference value.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F592

【参考文献】