摘要
提出了一种网络搜索信息提取及合成新技术——R/S-TDC-EMD-KPCA方法,首先利用重标极差法(R/S)和时差相关法(TDC)选择具有预测能力的关键词,并对关键词搜索量分别进行经验模态分解(EMD)降噪,然后利用核主成分(KPCA)方法合成网络综合搜索指数.以海南省月度游客流量为预测对象进行实证研究,结果表明,融合了网络综合搜索指数的模型在预测精度方面均优于其他基准模型.说明该文提出的方法能高质量的提取和合成网络搜索信息,进而可有效地应用于游客流量的辅助预测.
In this paper,a new web search information extraction and synthesis technology,the R/S-TDC-EMD-KPCA method,is proposed.First,the rescaled range method(R/S)and the time difference correlation(TDC)method are used to select the keywords with predictive ability.The empirical mode decomposition(EMD)noise reduction is performed on the selected keyword search volume,and then the kernel principal component(KPCA)method is used to synthesize the comprehensive network search index.An empirical study is conducted with the monthly tourist flow of Hainan Province as the forecast object.The results show that the model integrated with the network comprehensive search index is better than other benchmark models in terms of prediction accuracy.Therefore,the proposed method can extract and synthesize network search information with high quality,which can be effectively used for auxiliary prediction of tourist flow.
作者
曹静如
孙景云
Cao Jingru;Sun Jingyun(Lanzhou University of Finance and Economics;Center for Quantitative Analysis of Gansu Economic Development)
出处
《哈尔滨师范大学自然科学学报》
CAS
2022年第5期46-57,共12页
Natural Science Journal of Harbin Normal University
基金
国家自然科学基金项目(72061020)
甘肃省自然科学基金项目(21JR1RA280)
兰州财经大学2020年度高等教育教学改革研究重点项目(LJZ202008)
关键词
重标极差法
EMD降噪
KPCA
网络搜索信息
旅游预测
Rescaling range method
EMD denoising
KPCA
Internet search information
Tourism forecasting
作者简介
通讯作者:孙景云
