| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 402475 | Knowledge-Based Systems | 2012 | 10 Pages |
News recommendation is a very active research field. The number of online journals has increased in recent years owing to the increasing popularity of the Internet. In this context, it is important to offer user tools that facilitate faster and more accurate access to articles of interest in digital newspapers. We present two probabilistic models based on latent variables that recommend relevant news to users according to profiles of their visits to the newspaper website. As input, the models consider news content and categories, according to a predefined classification, of those news previously accessed. The experimental results show good performance with respect to baseline models in a data set of news extracted from a digital journal edition.
