Forming multidisciplinary teams is a key to carry out complex tasks, which is increasingly the case higher up in the knowledge value chain. Need for team recommendation systems has always been there, both in product companies or academy. Many studies show that in academic world number of authors per paper and coauthors per author are high and increasing. However finding the right collaborators is not an easy problem: usually people tend to work with the same set of personal acquaintances and miss new colleagues (and as a result, opportunities) as people are generally barely aware of experts and promising newcomers of the various topics involved to carry out a complex multidisciplinary work. In this talk, we present a study of this question as well as an implementation: T-RecS. We demonstrate that even if the question of finding the best team is probably irrelevant, some parameters seem to be worth investigating: expertise, cohesiveness and team repetitions. We list the various problems that a real world system has to cope with and present a framework for the same. Our implementation is limited to a specific dataset (NTU) but it can be extended to other domains subject to availability of suitable information. We eventually show that our system can leverage on social web applications to address some difficult sub-problems: sparsity and dirtiness of data, availability of potential collaborators, etc.
Nicolas a fait sa thèse avec Philippe. Il est invité par DRIM.
http://aventresque.free.fr/
