Timeline: A Dynamic Hierarchical Dirichlet Process Model for Recovering Birth/Death and Evolution of Topics in Text Stream

TL;DR

This paper introduces iDTM, a dynamic hierarchical Dirichlet process model that captures the birth, death, and evolution of topics over time in document streams, providing a comprehensive understanding of topic dynamics.

Contribution

The paper presents a novel infinite dynamic topic model that models all aspects of topic evolution, including birth, death, and gradual change, in temporal text streams.

Findings

Successfully models topic birth and death over time

Effectively captures topic evolution in real datasets

Outperforms existing models in simulated and real data

Abstract

Topic models have proven to be a useful tool for discovering latent structures in document collections. However, most document collections often come as temporal streams and thus several aspects of the latent structure such as the number of topics, the topics' distribution and popularity are time-evolving. Several models exist that model the evolution of some but not all of the above aspects. In this paper we introduce infinite dynamic topic models, iDTM, that can accommodate the evolution of all the aforementioned aspects. Our model assumes that documents are organized into epochs, where the documents within each epoch are exchangeable but the order between the documents is maintained across epochs. iDTM allows for unbounded number of topics: topics can die or be born at any epoch, and the representation of each topic can evolve according to a Markovian dynamics. We use iDTM to analyze the birth and evolution of topics in the NIPS community and evaluated the efficacy of our model on both simulated and real datasets with favorable outcome.