23. Juli deskWeb2.0: Combining Desktop and Social Search Sergej Zerr, Elena Demidova, Sergej Chernov L3S Research Center Hannover, Germany
Sergej Zerr2 Alice Desktop resources - Application form - communication Hotel recommendations from colleagues - Bookmarks Multimedia resources - Pictures - Videos News, people, links, etc. - Information about other conference participants Scenario: Trip Organization to SIGIR 2010
Sergej Zerr3 Alice Scenario: Desktop + Web 2.0 deskWeb2.0 - What are the integration challenges?
Sergej Zerr4 Integration Challenges Technical and semantic interoperability: Authentication, authorization, search, sharing Service specific data types, query syntax Ranking and aggregation of search results: Source relevance, quality Near duplicates Distance within the social network Diversification of results Privacy preserving resource sharing: Information about locations, people, events, routes X
Sergej Zerr5 Me MyFriend... FoF... Tree structure: Each user is represented as a node Direct friendship relations are edges Each user node can contain resources Goal: Obtaining a balanced result set over the tree
Sergej Zerr6 User Study: Tasks Question 1: Does search over social services contribute to desktop search with respect to the relevance of results? Question 2: Which types of desktop search such as location, people and general information finding benefit from social search and which do not? Types of desktop search: A) Find a person B) Find location information C) Find general information
Sergej Zerr7 Evaluation Types of desktop search: A) Find a person B) Find location information C) Find general information
Sergej Zerr8 Conclusions: Identification of challenges towards integrating desktop search with social search User study demonstrated an nDCG improvement after integration (person and general information search) Future research directions: Larger study Ranking and diversification of search results Using Web2.0 as a source for contextual information
Sergej Zerr9 Conclusions: Identification of challenges towards integrating desktop search with social search User study demonstrated an nDCG improvement after integration (person and general information search) Future research directions: Larger study Ranking and diversification of search results Using Web2.0 as a source for contextual information Thank you!