数据降维处理下旅游景区客流量长期预测研究  

Long Term Prediction Simulation of Tourist Flow in Tourist Attractions Based on Spatiotemporal Characteristics

作　　者：陈秀 任欣鹭 吴艳芳 CHEN Xiu;REN Xin-lu;WU Yan-fang(Guangzhou Huashang University,Guangdong Guangzhou 511300,China)

机构地区：[1]广州华商学院,广东广州511300

出　　处：《计算机仿真》2025年第2期514-518,共5页Computer Simulation

基　　金：2022年广东省教育厅认定科研项目(2022WQNCX096);广州华商学院校内导师制科研(2019HSDS27)。

摘　　要：旅游需求随时间变化,受季节性、天气条件、特殊事件(如节庆活动、体育赛事等)影响,客流量呈现出一种空间相关性以及非线性特征,难以挖掘密集区域客流量的变化规律,增加了旅游景区客流量预测的误差。因此,将客流量的时空特性考虑在内,提出一种数据降维处理下的旅游景区客流量长期预测新方法。将相邻水平与核密度作为景区客流量分布的时空特征,获取成片游览区、线性游览区的日环境流量,得到密集区域客流量的变化规律;引入拉格朗日乘子,对获取的客流量数据降维处理;利用遗传算法改进反向传播(Back Propagation,BP)神经网络,将降维的客流量数据作为输入,实现对旅游景区客流量预测。仿真结果表明,所提方法能捕捉到客流量变化规律,实现对未来较长一段时间的客流量预测,适应度在0.8以上,能够精确地捕捉客流量的变化,误差最小,且相关系数的值最趋近1。Tourism demand changes with time.Influenced by seasonality,weather conditions,and special events(such as festivals,sports events,etc.),the passenger flow shows a spatial correlation and nonlinear characteristics,which makes it difficult to explore the changing law of passenger flow in dense areas,increasing the error of passenger flow prediction in tourist attractions.Therefore,taking into account the spatio-temporal characteristics of tourist flow,a new method for long-term prediction of tourist flow in tourist attractions based on dimensionality reduction is proposed.Taking the adjacent level and core density as the spatial and temporal characteristics of the tourist flow distribution of the scenic spot,the daily environmental flow of the tourist areas in blocks and linear tourist areas can be obtained,and the change law of the passenger flow in dense areas can be obtained.A Lagrange multiplier is introduced to reduce the dimensions of the obtained passenger flow data.The genetic algorithm is used to improve the backpropagation(BP)neural network,and the reduced dimension passenger flow data is used as the input to realize the prediction of tourist flow in tourist attractions.The simulation results show that the proposed method can capture the change rule of passenger flow,realize the prediction of future passenger flow for a long period of time,and the fitness is more than O0.8.It can accurately capture the change of passenger flow,with the smallest error and the value of the correlation coefficient approaching 1.

关 键 词：时空特征 客流量预测 遗传算法 反向传播神经网络 数据降维 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]