Representation Learning with Depth and Breadth for Recommendation Using Multi-view Data

Abstract

Recommender system has been well investigated in the past years. However, the typical representative CF-like models often give recommendation with low accuracy when the interaction information between users and items are sparse. To address the practical issue, in this paper we develop a novel Representation Learning with Depth and Breadth (RLDB) model for better recommendation Specifically, we design a heterogeneous network embedding method and convolutional neural network based method to learn feature representations of users and items from user-item interaction structure and review texts, respectively. Furthermore, an end-to-end breadth learning model is proposed through employing multi-view machine technique to learn features and fuse these diverse types of features in a uniform framework. Extensive experiments clearly demonstrates that our model outperforms all the other methods in these datasets.