Attention with Long-Term Interval-Based Deep Sequential Learning for Recommendation

Modeling user behaviors as sequential learning provides key advantages in predicting future user actions, such as predicting the next product to purchase or the next song to listen to, for the purpose of personalized search and recommendation. Traditional methods for modeling sequential user behaviors usually depend on the premise of Markov processes, while recently recurrent neural networks (RNNs) have been adopted to leverage their power in modeling sequences. In this paper, we propose integrating attention mechanism into RNNs for better modeling sequential user behaviors. Specifically, we design a network featuring A ttention with L ong-term I nterval-based G ated R ecurrent U nits (ALI-GRU) to model temporal sequences of user actions. Compared to previous works, our network can exploit the information of temporal dimension extracted by time interval-based GRU in addition to normal GRU to encoding user actions and has a specially designed matrix-form attention function to characterize both long-term preferences and short-term intents of users, while the attention-weighted features are finally decoded to predict the next user action. We have performed experiments on two well-known public datasets as well as a huge dataset built from real-world data of one of the largest online shopping websites. Experimental results show that the proposed ALI-GRU achieves significant improvement compared to state-of-the-art RNN-based methods. ALI-GRU is also adopted in a real-world application and results of the online A/B test further demonstrate its practical value.