Construction And Application Of Knowledge Graph In Financial Field

The preservation and utilization of knowledge is the source of the continuation of human civilization for thousands of years.With the development of modern artificial intelligence industry,the emergence of knowledge graph has made the preservation and utilization of knowledge more diverse.The rise of the artificial intelligence industry is based on the condition of a large amount of regularized data and a reliable knowledge base.In recent years,with the development of machine learning and deep learning technology in the direction of natural language processing,the research and application of knowledge graphs have achieved great Progress.Constructing a knowledge graph using structured and unstructured text data involves a variety of natural language processing technologies such as knowledge extraction,knowledge fusion,and knowledge storage.Based on existing technologies and algorithms,this paper integrates financial data from different sources to build a knowledge graph that serves the financial industry.The main research contents of the article are as follows:(1)Research how to construct a financial knowledge graph,it mainly includes three parts: knowledge extraction,knowledge fusion and knowledge storage.Knowledge extraction uses Word2 Vec and BERT word vector models combined with Bi LSTM-CRF to form a knowledge extraction joint learning model,Knowledge fusion proposes an entity similarity calculation method(Leven-Cos Distance,LCD)based on the combination of edit distance and cosine on the basis of Leinstein's edit distance and word vector cosine distance,The top-n entities of similarity are selected from the entity candidate set to achieve knowledge fusion,and knowledge storage is stored in Neo4 j graph database.(2)Research how to implement intelligent question answering system.Question Semantic Parsing proposes a vector dictionary-based Bi LSTM-CRF entity recognition model and CNN relationship extraction model to identify entities in the question,through entity linking technology to disambiguate the entities in the question,and finally through Cypher query query The entity relationship in the sentence is searched in the knowledge base.(3)Based on(1)and(2),build a financial enterprise knowledge service platform.Using Python to design and develop a set of application platforms that integrate knowledge graph display,intelligent question answering,knowledge base update,authority management and other functions,complete the design and implementation of the entire system,including system requirements analysis,overall architecture,detailed design plan,and system implementation.This article aims to construct a knowledge graph of the financial field through financial enterprise information,and finally design and complete an enterprise information service platform based on the knowledge graph.