Research on intelligent recognition of data assets based on the MKG-JERE algorithm

The registration of data assets is a crucial measure for promoting the development of the data element market, contributing to the effective allocation of data resources. However, evaluation efforts face challenges due to the broad scope of knowledge required, low maturity of methods, and significant workload. This study addresses these challenges by proposing an intelligent recognition method for data assets based on a multi-layer knowledge graph-enhanced joint entity and relation extraction algorithm (MKG-JERE). This method enhances the accuracy and efficiency of data asset evaluation by constructing automatically generated lower-layer knowledge graphs that integrate system logs, corporate standard documents, and external knowledge, which are then refined to generate upper-layer knowledge graphs. Supported by multi-layer knowledge graphs, this method achieves automatic clustering of data resources, further constructing structured data products. Experimental results show that this method can reduce manual workload by at least 60%. The application prospects of this method are extensive, suitable for multiple fields, providing robust technical support for the efficient management of data assets.