Blog Annotation: From Corpus Analysis to Automatic Tag Suggestion

Nowadays, blogs cover a large audience and they become part of mainstream media.Tags and categories are structural elements of a blog post intended to increase a blog's visibility and enhance navigation and searching.We suppose that those annotations are made on subjective grounds rather than in a systematic way.This paper presents a 11 million words corpus of blogs posts in French dedicated to the analysis of blog post tagging and categorization practices.We present experiences on automatic tag and category suggestion based on this corpus.Preliminary results show that around 27% of the overall tags can be predicted from lexical frequency analysis of blog posts.Furthermore, a first comparison experience with an existing tag suggestion tool shows that an important proportion of the tags used for blog description are not present in the blog post.Preliminary observations of annotations in time might suggest that taking into account diachronic information from previous post might improve the tag suggestion process.