Hierarchical Attentive Transaction Embedding With Intra- and Inter-Transaction Dependencies for Next-Item Recommendation

A transaction-based recommender system (TBRS) aims to predict the next item by modeling dependencies in transactional data. Generally, two kinds of considered are intra-transaction dependence and inter-transaction dependence. Most existing TBRSs recommend only within current transaction while ignoring with recent transactions that may also affect item. However, as not all relevant items, ones should be identified prioritized. In this article, we propose a novel hierarchical attentive embedding (HATE) model tackle these issues. Specifically, two-level attention mechanism integrates both build an context representation incorporates intra- dependencies. With learned representation, HATE then recommends Experimental evaluations on real-world datasets show significantly outperforms state-of-the-art methods terms recommendation accuracy.