Computational Approaches to Analyzing Weblogs
Papers from the AAAI Spring Symposium
Nicolas Nicolov, Franco Salvetti, Mark Liberman, and James H. Martin, Cochairs
Weblogs are web pages that provide unedited, highly opinionated personal commentary including hyperlinks to other resources. Often, weblogs (also referred to as blogs) are chronological sequences of entries, maintained and published with authoring tools.
The “blogosphere” as a whole can be exploited for outreach opinion formation, maintaining online communities, supporting knowledge management within large global collaborative environments, monitoring the reaction to public events and is seen as the upcoming alternative to the mass media.
Semantic analysis of blogs represents the next challenge in the quest for understanding natural language. Their light content, fragmented topic structure, inconsistent grammar, and vulnerability to spam makes blog analysis extremely challenging when faced with questions like: Can the implicit and explicit communities implied by content and link structure be used to determine relevance and influence of bloggers? Can a blog segment be identified as a summary of a linked story in order to use both as training data for summarization research? Can we determine how information percolates through mass media outlets and blogs? Can blogs with multimedia content be stored in a way that allows search across different modalities? Can we find consumer complaints, discover vulnerabilities of products, and predict trends?
This symposium brought together researchers from different subject areas (such as computer science, linguistics, psychology, statistics, sociology, multimedia and the semantic web) and foster discussions about ongoing research.