Unraveling the dynamics of social governance innovation: A synergistic approach employing NLP and network analysis

Abstract

Society, inherently a complex network system, necessitates an equally intricate approach towards social governance. This study aims to dissect the spatio-temporal evolution of social governance innovation themes and explore the value network relationships among multiple stakeholders in social governance. To achieve this, we employ natural language processing to analyze text data from 285 social governance innovation cases. This analysis allows us to delve into the progression of social governance innovation practice themes, discern inherent patterns and modes, and reveal their spatio-temporal characteristics. These insights form the basis for our proposed recommendations on policy-making and public service delivery to address public problems. Concurrently, we utilize the ERNIE-UIE model for transfer learning to extract entities related to social governance stakeholders from the cases. Our findings suggest that governance practices extending beyond the government to include all possible stakeholders, such as enterprises and civil society, are essential. When combined with social network analysis, we can illuminate the value network relationships among these multiple social governance stakeholders surrounding particular innovative practice themes. Our study sheds light on the increasing importance of public–private partnerships and the emergence of social organizations as key agents of influence within various domains of social governance.