Fast, Flexible Models for Discovering Topic Correlation across Weakly-Related Collections
This addresses a problem for researchers analyzing large, asymmetric document collections in fields like sciences and humanities, though it is incremental as it builds on existing topic modeling methods.
The paper tackles the challenge of modeling weak topic correlations across document collections with varying numbers of topics, which are often overlooked by existing models, by introducing Correlated LDA and Correlated HDP, achieving efficient large-scale analysis on over 300k documents from JSTOR.
Weak topic correlation across document collections with different numbers of topics in individual collections presents challenges for existing cross-collection topic models. This paper introduces two probabilistic topic models, Correlated LDA (C-LDA) and Correlated HDP (C-HDP). These address problems that can arise when analyzing large, asymmetric, and potentially weakly-related collections. Topic correlations in weakly-related collections typically lie in the tail of the topic distribution, where they would be overlooked by models unable to fit large numbers of topics. To efficiently model this long tail for large-scale analysis, our models implement a parallel sampling algorithm based on the Metropolis-Hastings and alias methods (Yuan et al., 2015). The models are first evaluated on synthetic data, generated to simulate various collection-level asymmetries. We then present a case study of modeling over 300k documents in collections of sciences and humanities research from JSTOR.