基于奇异谱分析的旅游客流量组合预测模型  

The Combined Forecasting Mode of Tourist Passenger Flow Based on Singular Spectrum Analysis

作　　者：张娟 ZHANG Juan(Hulunbuir University,Hailar,Inner Mongolia 021008)

机构地区：[1]呼伦贝尔学院数学与统计学院,内蒙古海拉尔021008

出　　处：《呼伦贝尔学院学报》2022年第5期100-106,共7页Journal of Hulunbuir University

基　　金：2019年内蒙古自治区高等学校科学技术研究项目“呼伦贝尔旅游经济与生态环境的统计分析研究与应用”(NJZY19231);2022年呼伦贝尔学院科研资助项目“‘碳达峰’背景下呼伦贝尔市工业碳排放影响因素及预测研究”(2022FDYB03);2022年呼伦贝尔学院科研资助项目“基于神经网络的函数逼近算法研究”(2022ZKYB08)。

摘　　要：针对旅游客流量序列具有非平稳和非线性的特点,本文提出了一种SSA-WA加权平均组合预测模型。数据预处理阶段,采用奇异谱分析分解重构方法去除旅游客流量序列中的噪声;模型预测阶段,对去噪后的待预测序列分别采用HW方法、BP神经网络方法、LSSVM三个单一模型进行序列预测,并按照权重系数将三个单一模型的预测结果进行加权平均,得到最终的预测结果。以三亚市1988~2019年旅游客流量序列为例进行分析,通过与单一模型比较,发现本文建立的SSA-WA加权平均组合预测模型的预测误差更小,模型精度更高。Aiming at the non-stationary and nonlinear characteristics of tourism passenger flow series,this paper proposes the SSA-WA weighted average combined forecasting model.In the data preprocessing stage,the singular spectrum analysis decomposition and reconstruction method is used to remove the noise in the tourist flow sequence;In the model prediction stage,the HW method,the BP neural network method and the LSSVM three single models are used for the denoised sequence to be predicted.Sequence prediction,and weighted average of the prediction results of the three single models according to the weight coefficient to obtain the final prediction result.Taking the tourist passenger flow sequence in Sanya from 1988 to 2019 as an example,it is found that the prediction error of the SSA-WA combined mode established in this paper is smaller and the model accuracy is higher by comparing with a single mode.

关 键 词：旅游客流量 奇异谱分析 HW方法 BP神经网络方法 最小二乘支持向量机方法 组合预测 

分 类 号：O175.12[理学—数学]