NEG-MF:一种针对推荐系统的矩阵分解图嵌入模型  

NEG-MF,a matrix factorization graph embedding model for recommender systems

作　　者：赵素芬[1,2] Zhao Sufen(School of Computer Science,Central China Normal University,Wuhan,Hubei 430079,China;School of Computer Science,Wuhan University)

机构地区：[1]华中师范大学计算机学院,湖北武汉430079 [2]武汉大学计算机学院

出　　处：《计算机时代》2021年第9期6-9,共4页Computer Era

基　　金：华中师范大学中央高校基本科研业务费青年教师创新项目(No.CCNU20QN022)。

摘　　要：传统的矩阵分解图嵌入模型由于不对大量未知关系建模,其性能面临着很大的挑战性。为了提升矩阵分解模型的性能,提出了一种基于负采样技术的矩阵分解模型NEG-MF。该模型能够从跳数大于6的邻居节点中进行负采样,以降低模型生成图嵌入时对于负样本的偏差。在DBLP数据集上做的大量实验结果表明,相比其他的基线方法,基于NEG-MF的推荐算法在学术合作关系推荐问题上的性能有明显地提升。The traditional matrix factorization graph embedding model does not consider a large number of unknown relationships,so that its performance faces great challenges.In order to improve the performance of the generated embeddings,NEG-MF,a matrix factorization model based on negative sampling is proposed.When the model generates node embeddings,it can perform negative sampling from the neighbor nodes with hops>6 to reduce the bias of negative samples.A large number of experiment results on DBLP data sets show that,compared with the baseline methods,the performance of the recommendation algorithm based on the proposed NEG-MF has a significant improvement in the recommendation of academic collaborators.

关 键 词：矩阵分解 图嵌入 推荐系统 负采样 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]