Configurational Effects of Technology Convergence on Industrial Innovation from a Multidimensional Theoretical Perspective: A Tree-Model-Based Fuzzy-Set Qualitative Comparative Analysis

In the context of accelerating global technological and industrial transformation, we aim to systematically reveal the multifaceted patterns through which technology convergence drives industrial innovation. From the perspective of technology convergence, we develop a set of conditions shaping industrial innovation drawing on three theoretical lenses: micro-level technology recombination theory, meso-level co-evolution theory, and macro-level convergence chain theory. Following the innovation ecosystem framework, the analysis first employs the SHAP (SHapley Additive exPlanations) model to identify core variables. Subsequently, fsQCA (fuzzy-set Qualitative Comparative Analysis) is applied to derive configurational pathways, and the XGBoost (eXtreme Gradient Boosting) model is used to construct decision-tree pathways incorporating all variables. Finally, the two approaches are integrated to identify specific patterns through which technology convergence promotes industrial innovation. An empirical analysis of the biopharmaceutical sector identifies five distinct patterns that promote industrial innovation performance: institutional synergy–driven, network endogenous evolution, industry-driven innovation, demand-pull convergence, and disruptive innovation–driven patterns. These patterns constitute a multi-level synergistic framework. At the macro level, technology–industry coupling plays a dominant role. At the meso and micro levels, the external environment, collaboration among innovation actors, technological structure, and diffusion-related convergence factors are closely integrated. Through the coupling of intra-level and inter-level network elements, this framework jointly drives industrial innovation development.