Enhancing Team Composition in Professional Networks: Problem Definitions and Fast Solutions.

Area of Science:

• Computer Science
• Social Computing
• Organizational Behavior

Background:

• Effective team composition is crucial for collaborative environments.
• Existing methods for team member recommendations often overlook the interplay between team skills and social structure.
• A comprehensive approach is needed to maintain team performance during personnel changes.

Purpose of the Study:

• To develop a method for enhancing team composition by recommending members that minimize disruption to existing skills and social structures.
• To introduce a formalization of team-level similarity that integrates skill and structural aspects, along with their synergy.
• To propose computationally efficient algorithms for calculating this formalized team-level similarity.

Main Methods:

• Formalizing team-level similarity using graph kernels on attributed graphs to capture skill similarity, structural similarity, and their interaction.
• Developing fast algorithms by employing pruning strategies and exploring smoothness between team structures.
• Conducting extensive empirical evaluations on real-world datasets.

Main Results:

• The proposed method effectively recommends team members while preserving team performance.
• The integrated approach of skill and structural similarity, along with synergy, outperforms one-dimensional heuristic methods.
• The developed algorithms demonstrate significant efficiency in computation.

Conclusions:

• The formalized team-level similarity metric provides a comprehensive approach to team member recommendations.
• The proposed algorithms offer an effective and efficient solution for enhancing team composition in collaborative settings.
• This work contributes to optimizing team dynamics and performance in dynamic organizational environments.