Item Cold-Start Recommendation with Personalized Feature Selection

作     者：Yi-Fan Chen Xiang Zhao Jin-Yuan Liu Bin Ge Wei-Ming Zhang Yi-Fan Chen;Xiang Zhao;Jin-Yuan Liu;Bin Ge;Wei-Ming Zhang

作者机构：Science and Technology on Information Systems Engineering LaboratoryNational University of Defense TechnologyChangsha 410073China Academy of Military SciencesBeijing 100091China 

出 版 物：《Journal of Computer Science & Technology》 (计算机科学技术学报（英文版）)

年 卷 期：2020年第35卷第5期

页      面：1217-1230页

核心收录：

学科分类：080703[工学-动力机械及工程] 08[工学] 0807[工学-动力工程及工程热物理] 

基　　金：supported by the National Natural Science Foundation of China under Grant Nos.61872446,61902417,71690233,and 71971212 the Natural Science Foundation of Hunan Province of China under Grant No.2019JJ20024 

主　　题：high-dimensionality item cold-start top-TV recommendation personalized feature selection 

摘      要：The problem of recommending new items to users(often referred to as item cold-start recommendation)remains a challenge due to the absence of users’past preferences for these *** features from side information are typically leveraged to tackle the *** methods formulate regression methods,taking item features as input and user ratings as *** methods are confronted with the issue of overfitting when item features are high-dimensional,which greatly impedes the recommendation *** of high-dimensional item features,in this work,we opt for feature selection to solve the problem of recommending top-N new *** feature selection methods find a common set of features for all users,which fails to differentiate users1 preferences over item *** personalize feature selection,we propose to select item features discriminately for different *** study the personalization of feature selection at the level of the user or user *** fulfill the task by proposing two embedded feature selection *** process of personalized feature selection filters out the dimensions that are irrelevant to recommendations or unappealing to *** results on real-life datasets with high-dimensional side information reveal that the proposed method is effective in singling out features that are crucial to top-N recommendation and hence improving performance.