文献类型：期刊文献

英文题名：Multi-view heterogeneous fusion and embedding for categorical attributes on mixed data

作者：Li, Qiude Xiong, Qingyu Ji, Shengfen Gao, Min Yu, Yang Wu, Chao

第一作者：Li, Qiude

通信作者：Xiong, QY[1];Xiong, QY[2]

机构：[1]Chongqing Univ, Minist Educ, Key Lab Dependable Serv Comp Cyber Phys Soc, Chongqing, Peoples R China;[2]Chongqing Univ, Sch Big Data & Software Engn, Chongqing 400044, Peoples R China;[3]Guizhou Med Univ, Sch Biol & Engn, Guiyang 550004, Guizhou, Peoples R China;[4]Guizhou Inst Technol, Foreign Language Teaching Ctr, Guiyang 550003, Guizhou, Peoples R China

第一机构：Chongqing Univ, Minist Educ, Key Lab Dependable Serv Comp Cyber Phys Soc, Chongqing, Peoples R China

通信机构：corresponding author), Chongqing Univ, Minist Educ, Key Lab Dependable Serv Comp Cyber Phys Soc, Chongqing, Peoples R China;corresponding author), Chongqing Univ, Sch Big Data & Software Engn, Chongqing 400044, Peoples R China.

年份：2020

卷号：24

期号：14

起止页码：10843-10863

外文期刊名：SOFT COMPUTING

收录：;EI(收录号：20195007813721);Scopus(收录号：2-s2.0-85076095859);WOS:【SCI-EXPANDED(收录号:WOS:000500606300005)】；

基金：We thank anonymous reviewers for their valuable comments and suggestions. The work was supported by the Key Research Program of Chongqing Science & Technology Commission (Grant No. CSTC2017jcyjBX0025 and CSTC2019jscx-zdztzx0043), the Science and Technology Major Special Project of Guangxi (Grant No. GKAA17129002), the National Natural Science Foundations of China (Grant No. 61771077), and the National Key R&D Program of China (Grant No. 2018YFF0214706), Graduate Scientific Research and Innovation Foundation of Chongqing (Grant No. CYB19072 and CYS19028).

语种：英文

外文关键词：Categorical attributes; Coupling learning; Heterogeneous fusion; Metric learning; Embedding learning

摘要：Categorical attributes are ubiquitous in real-world collected data. However, such attributes lack a well-defined distance metric and cannot be directly manipulated per algebraic operations, so many data mining algorithms are unable to work directly on them. Learning an appropriate metric or an effective numerical embedding is very vital yet challenging, for categorical attributes with multi-view heterogeneous data characteristics. This paper proposes a novel multi-view heterogeneous fusion model (MVHF), which first captures basic coupling information for each view and then fuses these heterogeneous information from different views by multi-kernel metric learning, to measure the intrinsic distances between this type of categorical attributes; based on these measured distances, further, we use the manifold learning method to learn a high-quality numerical embedding for each categorical value. Experiments on 33 mixed data sets demonstrate that MVHF-enabled classification significantly enhances the performance, compared with state-of-the-art distance metrics or embedding competitors.