Neural Autoregressive Collaborative Filtering for Implicit Feedback

Abstract

This paper proposes implicit CF-NADE, a neural autoregressive model for collaborative filtering tasks using implicit feedback ( e.g. click, watch, browse behaviors). We first convert a users implicit feedback into a like vector and a confidence vector, and then model the probability of the like vector, weighted by the confidence vector. The training objective of implicit CF-NADE is to maximize a weighted negative log-likelihood. We test the performance of implicit CF-NADE on a dataset collected from a popular digital TV streaming service. More specifically, in the experiments, we describe how to convert watch counts into implicit relative rating, and feed into implicit CF-NADE. Then we compare the performance of implicit CF-NADE model with the popular implicit matrix factorization approach. Experimental results show that implicit CF-NADE significantly outperforms the baseline.

5 pages, 2 figures, accepted by DLRS2016 http://dlrs-workshop.org/

References in corpus (1)

• Distributed Representations of Words and Phrases and their Compositionality

Cited by in corpus (4)

• Deep Learning based Recommender System: A Survey and New Perspectives
• Model-Agnostic Counterfactual Reasoning for Eliminating Popularity Bias in Recommender System
• Graph Meta Network for Multi-Behavior Recommendation
• Multiplex Behavioral Relation Learning for Recommendation via Memory Augmented Transformer Network