Submitted on Nov 18, 2013


Tags : text_mining machine_learning generation

Author : David Blei, Associate Professor of Computer Science, Princeton University, U.S.A.


“Probabilistic topic models provide a suite of tohe documents. For example, consider readers clicking on articles in a newspaper website or scientists placing articles in their personal libraries
In this talk, I will review the basics of topic modeling and describe our recent research on collaborative topic models, which simultaneously analyze texts and corresponding user behavior data. We studied collaborative topic models on a large collection of 80,000 scientists’ libraries and the 250,000 abstracts of the corresponding articles. With this analysis, we can build recommendation systems that point scientists to articles they will like and, further, organize the scientific literature according to the discovered patterns of readership. As examples, we can identify articles that are important within a field and articles that transcend disciplinary boundaries.