The influence of service performance in China's sci-tech commissioner system: Using social network analysis and...

Area of Science:

• Agricultural Science
• Computer Science
• Sociology

Background:

• The Chinese government utilizes the Sci-Tech Commissioner System (SCS) to enhance agricultural development through science and technology.
• Assessing the performance of SCS is crucial for optimizing agricultural support and innovation.

Purpose of the Study:

• To categorize service group types within China's SCS using social network analysis (SNA) and machine learning (ML).
• To identify key factors influencing the service performance of sci-tech commissioners.
• To develop an interpretable ML model for understanding SCS dynamics.

Main Methods:

• Employed social network analysis (SNA) and clustering algorithms to categorize sci-tech commissioner groups.
• Utilized and compared various classification algorithms, selecting LightGBM for its accuracy.
• Applied SHAP (SHapley Additive exPlanations) to analyze factors impacting service performance.

Main Results:

• Identified distinct group types: small, cooperative young groups; larger, mixed-age groups; cooperative middle-aged/older groups; and isolated, influential individuals.
• Found that group structure and coordination ability are more critical to performance than group size.
• Highlighted the importance of good group structures and extensive social networks for high service performance.

Conclusions:

• Sci-tech commissioner service delivery is predominantly group-oriented.
• Optimizing group dynamics, structure, and inter-commissioner collaboration is essential for improving agricultural science and technology outreach.
• ML models, particularly LightGBM with SHAP analysis, provide valuable insights into complex service delivery systems.