Mining the Collaborative Networks: A Machine Learning-Based Approach to Firm Innovation in the Digital Transformation...

Area of Science:

• Business Administration
• Innovation Management
• Information Systems

Background:

• Rapid technological change necessitates understanding firm innovation drivers.
• Collaborative networks and digital transformation are critical factors influencing innovation.
• Existing research often overlooks the joint and nonlinear effects of these elements.

Purpose of the Study:

• To investigate the combined impact of collaborative network structures and digital transformation on firm innovation performance.
• To explore the nonlinear and interactive effects using a configurational perspective.
• To identify key network attributes and digital transformation configurations driving innovation.

Main Methods:

• Utilized patent data from Chinese listed manufacturing firms (2012-2022).
• Constructed inter-firm technological collaboration networks based on co-patenting.
• Employed Classification and Regression Tree (CART) model for configurational analysis, with regression robustness tests.

Main Results:

• Innovation performance is shaped by joint configurations of network attributes (structural holes, centrality) and digital transformation.
• Structural holes were identified as the most influential network determinant.
• Digital transformation interacts with network positions, leading to diverse innovation outcomes.

Conclusions:

• Firm innovation is a result of complex configurations, not isolated factors.
• A configurational approach, particularly using machine learning like CART, is superior for analyzing network-digital transformation interplay.
• Findings offer a novel framework for understanding collaborative innovation in the digital era.