基于网络搜索和CLSI-EMD-BP的旅游客流量预测研究

发布时间：2018-02-22 07:13

本文关键词： 网络搜索旅游预测 CLSI指数合成 EMD经验模态分解 BP神经网络　出处：《系统工程理论与实践》2017年01期 　论文类型：期刊论文

【摘要】：准确的旅游预测对于旅游政策制定当局和游客都具有重要意义,可以帮助资源的合理配置并避免拥堵事件和游客滞留事件的发生.为了提高旅游预测的准确性,本文考虑噪声在预测中的干扰,提出一种基于网络搜索的CLSI-EMD-BP预测模型.该模型首先利用CLSI方法对网络搜索数据进行指数合成,并利用EMD对序列进行噪声处理,将高频噪声从原序列中分离,再利用去噪处理后的网络搜索数据对旅游客流量进行预测.实证分析以九寨沟为例对预测期内未来22周旅游客流量进行预测发现,基于网络搜索的CLSI-EMD-BP预测误差显著低于时间序列、网络搜索和BP神经网络三个基准模型.该结论一方面说明了本文预测模型的改进作用,另一方面也表明了噪声处理在预测中的必要性.
[Abstract]:Accurate tourism prediction is of great significance for both tourism policy making authorities and tourists. It can help the rational allocation of resources and avoid the occurrence of congestion events and tourist stranded events, in order to improve the accuracy of tourism prediction. In this paper, considering the interference of noise in prediction, a CLSI-EMD-BP prediction model based on network search is proposed. Firstly, the CLSI method is used to synthesize the network search data exponentially, and the EMD is used to deal with the noise of the sequence. The high frequency noise is separated from the original sequence, and then the network search data after denoising is used to predict the tourist flow. The empirical analysis takes Jiuzhaigou as an example to predict the tourist flow in the next 22 weeks in the forecast period. The prediction error of CLSI-EMD-BP based on network search is significantly lower than that of time series, and the network search and BP neural network are three benchmark models. On the other hand, it also shows the necessity of noise processing in prediction.
【作者单位】：中国科学院大学经济与管理学院;中国科学院大数据挖掘与知识管理重点实验室;
【基金】：国家自然科学基金(71202115,71202155,71172199)~~
【分类号】：F592.7

【相似文献】