基于分类模型的日客流量预测  

Forecasting of Daily Visitor Numbers Based on the Classification Model

作　　者：马银超[1,2] 梁昌勇[1,2] 路彩红[1,2] 

机构地区：[1]合肥工业大学管理学院,安徽合肥230009 [2]合肥工业大学过程优化与智能决策教育部重点实验室,安徽合肥230009

出　　处：《国土资源科技管理》2015年第5期109-114,共6页Scientific and Technological Management of Land and Resources

基　　金：国家自然科学基金项目(71331002;71271072;71201045;71301037)

摘　　要：日客流量的准确预测能为景区管理决策提供科学可靠的依据。由于受到各种客观因素影响,日客流量不光呈现出强非线性特征,而且具有明显的季节性特征。整年度的日客流量数据跳跃波动过大,整年度模型很难对其进行准确预测。针对这一问题,根据黄山风景区日客流量的分布特点,建立了1、2和12月,3、6和11月,4和5月,7和8月,9和10月以及法定节假日六类预测模型,通过SVR预测方法的实证仿真显示,同年度模型相比基于客流量分布特征的分类模型明显消除了日客流量跳跃波动,有效地提高了预测精度。The accurate forecasting of daily visitor numbers can provide scientific and reliable basis for the decision-making management of scenic area.Due to various objective factors,the daily visitor numbers not only shows nonlinear characteristics but also has obvious seasonal characteristics.It is difficult to forecast daily visitor numbers accurately by the annual model because of the heavy fluctuation of annual daily data.In order to solve this problem,based on the distribution characteristics of daily visitor numbers of Huangshan scenic area,this paper established six classification forecasting models（January February and December,March June and November,April and May,July and August,September and October,official holiday）.Through simulation by SVR and compared with annual model,the results showed that six classification models eliminated the jump and fluctuations obviously,and improved the forecasting precision effectively.

关 键 词：日客流量预测 分类模型 SVR 黄山风景区 

分 类 号：F592[经济管理—旅游管理]