WANG Liwen SUN Wenfei LIU Meijing
New Economy. 2026, 47(4): 103-128.
Against the backdrop of the rapid development of the digital economy and its deep integration with the real economy, corporate digital transformation has become an important issue for improving resource allocation efficiency and competitive advantage. Existing studies on corporate digital transformation have mainly focused on single antecedent factors, paying limited attention to comparing firm characteristics, industry conditions, institutional environments, and other relevant factors within a unified analytical framework. This study integrates three dimensions—enterprise, industry, and institutional—to establish a comprehensive analytical framework for identifying prerequisite factors of digital transformation, overcoming the limitations of traditional linear models, and incorporates XGBoost, SHAP value, and partial dependence graphs. Using A-share listed companies in Shanghai and Shenzhen from 2010 to 2023 as the research sample, key predictive factors closely associated with corporate digital transformation are identified and interpreted within the framework, and their nonlinear relationships are visualizing subsequently.The main findings are as follows: first, compared with traditional linear regression models, machine learning models represented by XGBoost perform better in terms of out-of-sample goodness of fit and prediction error; second, R&D expenditure, fixed asset ratio, industry competition, industry size, legal institutional environment, firm size, asset-intensive industry status, media attention, customer concentration, and regional industrial structure upgrading are relatively important drivers of corporate digital transformation; third, firm age, customer concentration, supplier concentration, media attention, and analyst coverage exhibit nonlinear relationships with digital transformation. By applying machine learning methods, this study provides a more comprehensive perspective for research on digital transformation it also provides theoretical foundations and practical insights for enterprises to optimize resource allocation and for the government to improve the institutional environment.