Personalizing (Re-ranking) Web Search Results Using Information Present On A Social Network
Abstract
We describe a social search engine paradigm which can be built on top of a classic search engine (e.g. Google, Yahoo, etc.) and a social information network (such as FriendSter). In this thesis, the objective was to design algorithms and develop methods to efficiently combine information available in the underlying systems (Search Engine & Social Information Network) to better satisfy the search needs of a user. We are interested on how to efficiently employ social information to re-order a list of URLs retrieved by querying a search engine. The objective was to re-order the list of URLs in a way that favors URLs that are more relevant to user's interest towards a personalized search engine. We conduct a thorough user study & come up with certain experiments to show some of the functionality of the system.