Dynamic Collaborative Filtering with Compound Poisson Factorization

TL;DR

This paper introduces a dynamic matrix factorization model that captures the evolving nature of user preferences and item perceptions over time, improving prediction accuracy in collaborative filtering tasks.

Contribution

It presents a novel conjugate, numerically stable dynamic matrix factorization model based on compound Poisson processes and Gamma-Markov chains, with a stochastic variational inference method.

Findings

DCPF outperforms static models in predictive accuracy

Model effectively captures drifting latent factors over time

Applied successfully to Netflix, Yelp, and Last.fm datasets

Abstract

Model-based collaborative filtering analyzes user-item interactions to infer latent factors that represent user preferences and item characteristics in order to predict future interactions. Most collaborative filtering algorithms assume that these latent factors are static, although it has been shown that user preferences and item perceptions drift over time. In this paper, we propose a conjugate and numerically stable dynamic matrix factorization (DCPF) based on compound Poisson matrix factorization that models the smoothly drifting latent factors using Gamma-Markov chains. We propose a numerically stable Gamma chain construction, and then present a stochastic variational inference approach to estimate the parameters of our model. We apply our model to time-stamped ratings data sets: Netflix, Yelp, and Last.fm, where DCPF achieves a higher predictive accuracy than state-of-the-art static and dynamic factorization models.