Abstract:     The data element has become an important production factor in promoting the development of the digital economy, and studying the utilization level of data elements is of great significance for advancing the growth of the digital economy. This paper, based on the Technology-Organization-Environment (TOE) theoretical framework, adopting the fuzzy-set qualitative comparative analysis method (fsQCA) and artificial neural network (ANN) approaches, introduced the interpretable machine learning model SHAP to investigate the influencing factors of data element utilization levels. Using the data from 30 provinces (autonomous regions and municipalities) in China, this study empirically analyzed the factors influencing data element utilization levels. The research found that the synergy among technological, organizational and environmental factors forms three configurational paths capable of achieving high data element utilization levels: innovation-organization-openness composite, technology-organization-environment composite, and infrastructure-institution-environment composite. Among these, technological innovation levels and data resource openness have the most significant impact on achieving high data element utilization levels, while a singular focus on data resource openness may have a negative effect on the utilization of data elements. The research conclusions have clarified the influencing factors and configurational paths of data element utilization levels at the theoretical level, enriched the relevant studies on data elements, and proposed a hybrid fsQCAANN method, thus offering a new perspective for addressing similar complex configurational problems. At the practical level, the findings will provide some theoretical support and managerial insights for optimizing data element utilization policies and enhancing data element utilization levels in China.

Key words: data element, digital economy, machine learning, artificial neural network, fsQCA, SHAP