A Framework of Online Community based Expertise Information Retrieval on Grid Status of This Document

Web-based online communities such as blogs, forums and scientific communities have become important places for people to seek and share expertise. Search engines such as Google, Yahoo!, Live etc. are not yet capable to address queries that require deep semantic understanding of the query or the document. Instead, it may be preferable to find and ask someone who has related expertise or experience on a topic. Web-based online communities are the places people often seek advice or help. Before an analysis of search capabilities for these communities can be done, we need to gather the data (questions and answers, social support or discussion, comments or advice, content rating, social relations, and so forth) that describe the communities. There is no universal standard data structure for the outline of user participation in these communities. Also, as these communities rarely interoperate, each typically only has access to its own social data and cannot benefit from other communities’ data. Extracting, aggregating and analyzing data from these communities for finding experts on a single framework is a challenging task. In this document, we present a Grid-enabled framework of expertise search (GREFES) engine, which utilizes online communities as sources for experts on various topics. We suggest an open data structure called SNML (Social Network Markup Language) to outline user participation in online communities. The architecture addresses major challenges in crawling of community data and query processing by utilizing the computational power and high bandwidth inherently available in the Grid. Our framework supports open APIs for third party providers or developers to build new solutions in order to get more user feedback to improve the system. GFD-I.164 January 14, 2010