基于情感分析的旅游客流量多尺度组合预测模型  

Multi-scale Combined Prediction Model of Tourist Flow Based on Sentiment Analysis

作　　者：刘金培[1] 代玉洁 陈意 Liu Jinpei;Dai Yujie;Chen Yi(School of Business,Anhui University,Hefei 230601,China)

机构地区：[1]安徽大学商学院,合肥230601

出　　处：《统计与决策》2022年第22期29-34,共6页Statistics & Decision

基　　金：基金资助项目(2008085MG226);安徽省高校人文社会科学重点研究项目(SK2020A0054)。

摘　　要：文章提出了一种基于在线文本情感分析的旅游客流量多尺度组合预测模型。首先,用Python爬取客流量历史数据和旅游网站的评论并对其进行预处理,并使用Snownlp情感分析计算处理后的评论情感值作为客流量的影响因素;其次,将客流量和评论情感值分别用互补集成经验模态分解分解为不同的本征模函数,并用样本熵将其重构为高频、低频和趋势序列;然后,用反向传播神经网络、支持向量机和长短期记忆神经网络模型分别对高频、低频和趋势序列进行预测;最后,把三种方法的预测值相加即可获得最终预测结果。以四姑娘山风景区为例进行实证分析,结果显示所提模型能很好地提高客流量的预测效果,具有应用价值。This paper proposes a combined prediction model of tourist flow based on online text sentiment analysis.Firstly,Python is employed to crawl the historical data of tourist flow and reviews of travel websites and pre-process them;Snownlp sentiment analysis is used to calculate the sentiment value of the comments after processing as the influencing factor of passenger flow.Secondly,empirical mode decomposition of complementary integration is used to decompose tourist flow and comment sentiment value into different intrinsic mode functions,and they are reconstructed into high-frequency sequence,low-frequency sequence and trend sequence by using sample entropy.Then,back-propagation neural network,support vector machine and long short-term memory neural network model are used to predict the high-frequency sequence,low-frequency sequence and trend sequence,respectively.Finally,the prediction results of the three prediction methods are summed to obtain the final predicted result.The empirical analysis of Siguniang Mountain Scenic Area shows that the proposed model can improve the performance of tourist flow prediction and has good application value.

关 键 词：情感分析 CEEMD 支持向量机 长短期记忆神经网络 

分 类 号：O212[理学—概率论与数理统计]