基于整体序列建模的会话推荐模型

Session-based recommendation model based on overall sequence modeling

会话推荐常用的循环神经网络根据会话的短期交互生成相关表示。循环神经网络采用多次迭代的方式逐步生成表示,在每次迭代中选取局部最优解,无法从全局角度考虑会话记录中各项目的重要性,使会话表示的性能受限。针对该问题,该文提出一种基于整体序列建模的会话推荐模型,通过综合考虑各项目内容对会话表示的重要性,生成更为有效的会话表示。为了更好地挖掘项目内容在会话推荐中的重要性,该文联合分析了该项目在会话历史中所处的位置和该项目与会话最新交互项目的关系等信息,衡量该项目在会话中的权重,最后融合会话历史中的各项目信息生成整体的会话表示。在3个公开数据集上进行的大量实验表明,该文方法明显优于现有方法,证明了其有效性。

Recurrent neural network based on session-based recommendation mainly explores the short-term interactions of sessions to generate the derived representations.Recurrent neural network progressively generates the representation through multiple iterations,and selects the local optimal solution in each iteration.It is impossible to consider the importance of each item in the session from a global perspective,which limits the effectiveness of the session representation.To solve this problem,this paper proposes a session-based recommendation model based on overall sequence modeling.By exploring the importance of each item to the session representation as a whole,a more effective session representation is learned.To well uncover the importance of each item for the session-based recommendation,this paper jointly explores the position of the item in the history session as well as the relationship between the item and the latest item in the session.The desired session representation is obtained by fusing the information of the items in the history session.Extensive experiments conducted on three public data sets show that the method in the paper is significantly better than the state-of-the-art methods.