Deep Learning-Based Named Entity Recognition and Knowledge Graph for Accidents of Commercial Bank

Abstract

With the diversified development of business, the construction of the banking system has become increasingly complex, which is prone to accidents. Since system accidents are the result of the combined action of various risk factors, accident management requires comprehensive knowledge support. Although bank accident management has accumulated a large amount of data, there is still a lack of effective solutions to obtain the required knowledge from big data quickly and accurately when faced with a specific accident. To solve the above problems, we developed bank accident management from the perspective of knowledge support to introduce relevant methods and technologies in the field of artificial intelligence. Then, accident management based on named entity recognition and knowledge graph can be developed. The entity annotation corpus in banking accidents is constructed. For the context of each bank accident, key information (four types of entities: time, accident name, loss amount, and reason) is automatically extracted by the BERT-BiLSTM-CRF model. Various entities and relational knowledge elements in the knowledge graph are retained in the graph database Neo4j to form a knowledge graph in the field of banking accidents. We provide important references for the bank's accident analysis, cause investigation, resource allocation, and management decision-making.