TPGN: A Time-Preference Gate Network for e-commerce purchase intention recognition

Abstract The studies on users’ purchase intentions based on e-commerce data are of great significance to marketers, buyers, and society. Current studies on users’ intentions with traditional machine learning methods usually focus on unique features and are time-consuming. Due to the characteristics of user behaviors and the importance of time sequence, deep learning methods are increasingly applied in relevant studies. In the study, in order to predict online user’s purchase intentions, based on Long-Short Term Memory (LSTM) model, we proposed Time-Preference Gate Network (TPGN). A pair of preference gates and a pair of time interval gates are added to the model. The preference gates are used to capture the users’ category preferences at different time and the time interval gates are used to capture the users’ long-term interest. In our model, through coupling the input gate with the forget gate, the parameters of the model are reduced and the performance is improved. In addition, the difference in user gender leads to the difference in the performance. Extended experiments with two real data sets confirmed that our model performed better than other baseline models.