Greedy-Based Precise Expansion Algorithm for Customized Group Team Formation Problem

Abstract

The team formation problem is a social network research issue. Studies on this topic have recently proposed various team structures. However, these structures have been inflexible and imperfect. Moreover, in order to build teams with the lowest communication cost, proposed structures have always contained some communicators assigned to the team who lack expert skills. To improve upon defective structures, we propose the Customized Group Team Formation. Teams can be constructed automatically without any predefined structure. This allows project leaders to choose exemplary teammates based on skill requirements. Additionally, the proposed greedy-based precise expansion algorithm, the method of expertise query closest to the real world, can also save preprocessing time and reduce the amount of non-professional people on the team. Experiments on real datasets show effective performance on communication cost and cardinality.