基于ARIMA与SVM组合模型的国内旅游市场预测研究

发布时间：2018-01-16 03:27

本文关键词：基于ARIMA与SVM组合模型的国内旅游市场预测研究　出处：《东华理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：改革开放以来,随着中国经济与国民收入的增长,旅游业这项活动日益大众化,并在政治、经济、社会、文化、生态等领域显示出巨大活力。目前,中国国内旅游已成为世界上增速最快、数量最多、潜力最大的旅游市场,其发展规模已超过入境旅游和出境旅游,在经济发展中占据极其重要的地位。因此,如何在未来中把握和预测国内旅游市场发展的趋势,为旅游企业和管理者做出正确决策就成为政府必须要解决的现实问题。本文首先对国内旅游市场的背景进行了探讨,论述了研究的内容、方法和意义。接着对1995-2015年影响国内旅游市场因素的数据进行了灰色关联度分析,通过平均绝对百分误差(MAPE)的大小来判断各个模型对国内旅游人数预测的精度。在此基础上,对交通和城市居民人均可支配收入进行预测并利用预测数据对国内旅游人数进行预测。最后根据未来国内旅游人数的持续增长所带来的问题(交通拥挤、旅游景区安全、旅游服务质量下降和市场秩序混乱)提出一些建议和对策。主要研究结论如下:(1)运用灰色关联分析法对影响国内旅游市场的因素(交通(公路铁路总里程)、旅游产品价格(CPI)、旅游环境(国内旅游收入)和城市居民人均可支配收入)进行关联度分析,结果证明交通对其影响最大,其次是城市居民人均可支配收入。(2)运用差分自回归移动平均(ARIMA)、支持向量机(单因素SVM和多因素SVM)以及组合模型对国内旅游人数进行预测比较,并考虑到国内旅游人数时间序列数据的线性与非线性的特征,结果证明组合模型预测的精确度更高,泛化能力更强。(3)全国国内旅游人数与交通(公路铁路总里程)、旅游产品价格(CPI)、旅游环境(国内旅游收入)和城市居民人均可支配收入存在高度正相关,特别是在经济发展的初期,经济发展对国内旅游人数的依赖性较强,并运用泛化能力更强、精确度更高的组合模型对影响国内旅游人数最大的两个因素(交通和可支配收入)预测,然后通过三次多项式回归对其进行预测,结果发现,交通对国内旅游人数的预测结果更为可靠。(4)全国国内旅游人数在1995-2004年增加比较平稳,2005-2015年旅游人数增加趋势加速。从预测结果来看,未来十年国内旅游人数会趋增,旅游市场规模会进一步扩大,人均出游率达9.6次,旅游消费将大大提高。这和国家对旅游产业的大力支持及经济结构的调整和转型是相吻合的。
[Abstract]:Since the reform and opening up, with the growth of China's economy and national income, tourism has become increasingly popular, and has shown great vitality in political, economic, social, cultural, ecological and other fields. China's domestic tourism has become the world's fastest growing, the largest number, the largest potential tourism market, its development scale has exceeded the inbound tourism and outbound tourism, occupies an extremely important position in the economic development. How to grasp and predict the development trend of domestic tourism market in the future. To make the correct decision for tourism enterprises and managers becomes a realistic problem that must be solved by the government. Firstly, this paper discusses the background of domestic tourism market and discusses the content of the research. Methods and significance. Then the data of influencing factors of domestic tourism market from 1995 to 2015 were analyzed by grey correlation degree. Through the average absolute percent error of the size of MAPE to judge the accuracy of each model for the domestic tourist population prediction. On this basis. Forecast the per capita disposable income of traffic and urban residents and forecast the number of domestic tourism by using the forecast data. Finally, according to the problems caused by the sustained growth of the number of domestic tourism (traffic congestion) in the future. Scenic spots are safe. Some suggestions and countermeasures are put forward for the decline of tourism service quality and confusion of market order. The main conclusions are as follows: 1) the grey relational analysis is used to analyze the factors (traffic) affecting the domestic tourism market. Total road and rail mileage). The correlation analysis of tourism product price CPI, tourism environment (domestic tourism income) and per capita disposable income of urban residents shows that traffic has the greatest impact on it. The second is the per capita disposable income of urban residents. 2) using differential autoregressive moving average (ARIMA). Support vector machine (single factor SVM and multivariate SVM) and combination model are used to predict and compare the number of domestic tourists, taking into account the linear and nonlinear characteristics of the time series data of domestic tourist population. The results show that the forecasting accuracy of the combined model is higher and the generalization ability is stronger. 3) the number of domestic tourism and transportation (total mileage of road and railway, the price of tourism product is higher than CPI). Tourism environment (domestic tourism income) and per capita disposable income of urban residents have a high positive correlation, especially in the early stage of economic development, economic development has a strong dependence on the number of domestic tourism. And using the combination model with stronger generalization ability and higher accuracy to predict the two factors (traffic and disposable income) that affect the number of domestic tourism, and then predict it by cubic polynomial regression. The result shows that the forecast result of the traffic to the domestic tourist population is more reliable. 4) the increase of the domestic tourist population in the whole country from 1995 to 2004 is relatively stable. From the forecast results, the number of domestic tourism will increase in the next decade, the scale of the tourism market will further expand, the per capita travel rate reaches 9.6. Tourism consumption will be greatly increased. This is consistent with the state's strong support for the tourism industry and the adjustment and transformation of the economic structure.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F592.6

【参考文献】