Proceedings of the Workshop on Analytics for Noisy Unstructured Text Data, 20th International Joint Conference on Artificial Intelligence (IJCAI-2007)
BlogVox: Separating Blog Wheat from Blog Chaff
January 7, 2007
Blog posts are often informally written, poorly structured, rife with spelling and grammatical errors, and feature non-traditional content. These characteristics make them difficult to process with standard language analysis tools. Performing linguistic analysis on blogs is plagued by two additional problems: (i) the presence of spam blogs and spam comments and (ii) extraneous non-content including blog-rolls, link-rolls, advertisements and sidebars. We describe techniques designed to eliminate noisy blog data developed as part of the BlogVox system - a blog analytics engine we developed for the 2006 TREC Blog Track. The findings in this paper underscore the importance of removing spurious content from blog collections.
Downloads: 3058 downloads
Google Scholar Citations: 8 citations