Modeling personalized representation for within-basket recommendation based on deep learning

Abstract

Within-basket recommendation, which predicts the related item to be added to the basket from the item corpus, is prevalent in grocery shopping and e-commerce. Besides user-item collaborative filtering information, the retail platform also needs to combine the items that the user currently owns to make a recommendation. Previous work solves the task by rule mining or incorporating various types of associations. However, the representation of the basket and the high-order feature interaction is hardly investigated previously. In this work, we propose a deep learning-based model named DBFM (Deep Basket-Sensitive Factorization Machine) to address the task. We first make a personalized representation for a basket based on its constituent items instead of ID by latent factor learning, which improves the generalization of the model to baskets. Then we combine both low-order and high-order feature patterns to capture the sophisticated structures from inputs. Finally, a linear function is used to integrate with the results of different components. Experiments on three real-world datasets demonstrate higher performance of our model over state-of-the-art methods.