Entity Extraction for Dispatching Automation Job Based on Knowledge Base

With the development of the electric power industry, the field of dispatch automation system operation and maintenance has accumulated a huge amount of unstructured data, and the historical data is difficult to be transformed into a priori knowledge, so the process of digital transformation in the field of power dispatch automation system operation and maintenance is slow. The use of knowledge graph to integrate multiple structural forms of data in power dispatch automation system helps to realize intelligent control and support the automation and intelligent transformation of dispatching operations. Entity extraction aims to extract the relationship between entities from massive and heterogeneous data, which is a key step in the construction of knowledge graph. In this paper, we propose an auxiliary method of entity extraction for dispatching automation job. Firstly, for the multimodal dispatching automation system's operation and maintenance data, text information was extracted through encoding, and text data preprocessing was conducted by combining the specific professional operation and management guidelines of power dispatching automation with the distinct data characteristics inherent to the system operation and maintenance domain; subsequently, text data redundancy was reduced through similarity computation, and ultimately, entities were extracted employing the Latent Dirichlet Allocation (LDA) algorithm to establish an entity knowledge base. Experiments show that the entity extraction knowledge base constructed in this paper can effectively improve the accuracy of entity extraction tasks.