Abstract (EN):
In this paper we describe an application of recommender systems to team building in a company or organization. The recommender system uses a collaborative filtering model based approach. Recommender models are sets of association rules extracted from the activity log of employees assigned to projects or tasks. Recommendation is performed at two levels: first by recommending a single team element given a partially built team; and second by recommending changes to a completed team. The methodology is applied to a case study with real data. The results are evaluated through experimental tests and one survey to potential users.
Language:
English
Type (Professor's evaluation):
Scientific