User intent transition for explicit collaborative search through groups recommendation

Abstract

Collaborative information retrieval is an emerging research field in charge of establishing techniques and methods to satisfy the shared information needs of groups of people that work together as a team, starting from the extension of the information seeking and retrieval process with the knowledge about the queries, the context, and the explicit collaboration habits among them. Unfortunately, in broad online communities that besides grow continuously (e.g., social networks, e-learning systems, or peer-to-peer networks) can be difficult the conformation of these groups, or all benefits can not be obtained, for the lack of transparency among users' seeking tasks in distributed environments. To address this issue, we propose in this work a recommender agent based on latent semantic indexing formalism to assist the users that search alone to find and join to groups with similar information needs. With this mechanism, a user can change easily her solo search intent to explicit collaborative search. We assume as hypothesis that both, the group and the new member will benefit. To validate our hypothesis we have designed an experiment with twelve groups of students in the context of search-driven software development.