基于全局采样的混合门控神经网络推荐模型  

Hybrid gated neural network recommendation model based on global sampling

作　　者：张意灵 邵雄凯[1] 高榕 王春枝[1] 吴歆韵 ZHANG Yi-ling;SHAO Xiong-kai;GAO Rong;WANG Chun-zhi;WU Xin-yun(School of Computer Science,Hubei University of Technology,Wuhan 430068,China;State Key Laboratory for Novel Software Technology,Nanjing University,Nanjing 210093,China)

机构地区：[1]湖北工业大学计算机学院,湖北武汉430068 [2]南京大学计算机软件新技术国家重点实验室,江苏南京210093

出　　处：《计算机工程与设计》2022年第12期3365-3372,共8页Computer Engineering and Design

基　　金：国家自然科学基金项目(61170135、61902116);南京大学计算机软件新技术国家重点实验室开放课题基金项目(KFKT2021B12);湖北省高层次人才基金项目(GCRC2020011);湖北工业大学博士科研启动基金项目(BSQD2019026、BSQD2019022)。

摘　　要：为缓解在数据稀疏情况下,推荐性能降低的问题,提出一个基于融合评分数据和文本数据的混合推荐模型。采用词级注意模块从项目内容描述和用户评分中学习项目的隐藏特征,使用一种门控神经网络对获得的隐藏特征进行更有效的特征融合,采用集成邻域注意力和自动编码器的学习框架对用户的偏好进行建模。采用基于新的全局采样策略优化相关损失函数,提供更加精准的推荐服务。实验结果表明,该推荐模型优于主流先进推荐模型。To solve the problem of degraded recommendation performance in the presence of sparse data,a hybrid recommendation model based on fused rating data and text data was proposed.A word-level attention module was used to learn hidden features of items from item content descriptions and user ratings,a gated neural network was used to perform more effective feature fusion on the obtained hidden features,and a learning framework with integrated neighborhood attention and autoencoder was used to model users’preferences.A new global sampling strategy based on a new global sampling strategy was used to optimize the correlation loss function to provide more accurate recommendation services.Experimental results show that the proposed recommendation model outperforms the mainstream advanced recommendation model.

关 键 词：推荐系统 多源信息 融合策略 隐式反馈 门控神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]