Dynamic topic models
David M. Blei, John Lafferty · International Conference on Machine Learning · 2006
A family of probabilistic time series models is developed to analyze the time evolution of topics in large document collections. The approach is to use state space models on the natural parameters of the multinomial distributions that represent the topics. Variational approximations based on Kalman filters and nonparametric wavelet regression are developed to carry out approximate posterior inference over the latent topics. In addition to giving quantitative, predictive models of a sequential corpus, dynamic topic models provide a qualitative window into the contents of a large document collection. The models are demonstrated by analyzing the OCR'ed archives of the journal Science from 1880 through 2000.
Cited by 2,386 works. doi:10.1145/1143844.1143859
Authors
References
- A New Approach to Linear Filtering and Prediction Problems
- Latent dirichlet allocation
- Inference of Population Structure Using Multilocus Genotype Data
- A Bayesian Hierarchical Model for Learning Natural Scene Categories
- Finding scientific topics
- Discovering objects and their location in images
- Bayesian Forecasting and Dynamic Models
- The author-topic model for authors and documents
- The Statistical Analysis of Compositional Data
- Sparse Gaussian Processes using Pseudo-inputs