文献类型：期刊文献

英文题名：Lean improvement of the stage shows in theme park based on consumer preferences correlation deep mining

作者：Li, Shugang[1];Lu, Hanyu[1];Kong, Jiali[1,2];Yu, Zhaoxu[3];Wang, Ru[1]

机构：[1]Shanghai Univ, Sch Management, Shanghai 200444, Peoples R China;[2]Ind & Commercial Bank China, Shanghai Branch, Shanghai 200131, Peoples R China;[3]East China Univ Sci & Technol, Dept Automat, Shanghai 200237, Peoples R China

年份：2020

卷号：79

期号：33-34

起止页码：24487

外文期刊名：MULTIMEDIA TOOLS AND APPLICATIONS

收录：;EI(收录号：20202608867388);WOS:【SSCI(收录号:WOS:000542143000004)，SCI-EXPANDED(收录号:WOS:000542143000004)】；

基金：This work was supported by the Chinese National Natural Science Foundation (No. 71871135).

语种：英文

外文关键词：Consumer preferences; Correlation deep mining; Kano-IPA model; Product improvement; Theme park

摘要：Online comments provide a new and convenient way to understand consumer preference, but these comments for stage shows in theme park are usually incomplete, which can seriously affect the accuracy of existing mining models. In order to overcome the dilemma of missing information, we propose the consumer preferences correlation deep mining model, which precisely mines user preferences from two aspects: comment semantic deep mining and attribute emotion correlation mining. Furthermore, the Kano-IPA model is proposed to comprehensively excavate the user satisfaction and the importance of product attributes to give a lean improvement strategy for stage shows. Specifically, firstly, correlation deep mining model is constructed to deeply mine the missing attribute emotional polarity based on the emotional correlation sequence, emotional vector and Senti2vec + Gated Recurrent Unit model. Secondly, correlation width mining model is developed to excavate the user preferences for the stage shows attribute. In the correlation width mining model, the partial regression equation is used to describe the influence of the user emotional polarity on the user satisfaction level. Based on the emotion correlated attribute sequences, the correlation Kano mapping rules are proposed, and then the priority of user preferences for product attributes is given. Thirdly, the Kano-IPA model is designed for the lean improvement of products to achieve higher benefits at a lower cost. Finally, the experimental results on Shanghai Disneyland confirm the effectiveness and application value of the proposed model. Consequently, this study provides an accurate decision support model driven by big data for product improvement.