摘要
随着首个在线旅游数据生态共建倡议书的发布,在线评论数据更加真实、准确地表达顾客的客观感受,成为商家和消费者情报的重要来源。结合LDA、TF-IDF算法获取不同类型酒店客户评论特征权值,采用AipNLP获得情感倾向性估计值。利用Lasso算法进行特征筛选构建基于Lasso-LDA的用户偏好模型,将该模型应用于携程网上五种类型用户的偏好分析中。研究结果表明,与传统的多元线性回归及岭回归相比,该模型有更好的预测效果。
With the publication of the first online tourism data ecology co-construction proposal,online commentary data more truthfully and accurately express customers objective feelings,and become an important source of business and consumer information.This paper combined LDA and TF-IDF algorithms to obtain the feature weights of different types of hotel customer reviews,then used AipNLP to obtain the emotional orientation estimates.The Lasso algorithm was used to filter the features and construct a user preference model based on Lasso-LDA,which was applied to the preference analysis of 5 types of users on Ctrip.The results show that,compared with the traditional multiple linear regression and ridge regression,the hotel user preference model constructed in this paper has better prediction effect.
作者
赵志杰
刘岩
张艳荣
周婉婷
孟令跃
Zhao Zhijie;Liu Yan;Zhang Yanrong;Zhou Wanting;Meng Lingyue(School of Computer and Information Engineering,Harbin University of Commerce,Harbin 150028,Heilongjiang,China;Heilongjiang Key Laboratory of Electronic Commerce and Information Processing,Harbin University of Commerce,Harbin 150028,Heilongjiang,China)
出处
《计算机应用与软件》
北大核心
2021年第2期19-26,共8页
Computer Applications and Software
基金
教育部人文社会科学研究项目(18YJAZH128)。
