The Design And Implementation Of Financial Knowledge Graph Based On Information Extraction Technology

With the continuous development of economy and society,the demand for information in all walks of life is becoming more and more urgent,especially in the financial field.Domain information is helpful to improve the financial company's business,including company background survey,stock market trading strategy formulation and so on.However,most of the information in the financial field is released through company announcements,research reports and other forms,requiring professionals to read a large number of documents,resulting in waste of human resources.Information Extraction is a very important branch of natural language processing.It mainly solves the problem of extracting event information from semi-structured or unstructured data.It has great practical significance to use this technology to extract financial announcement information text.Financial announcement text data has very obvious characteristics.Usually,a financial announcement text will express a financial event in a standard and concise language.In this event sentence,the current business activities of financial companies are disclosed.Because of the high demand for information accuracy in the financial field and the lack of large-scale annotated data,it is difficult to extract information from such texts.At the same time,after extracting information,how to use information reasonably is the key.The main contributions of this paper are as follows:1.This paper studies the event extraction of financial announcement text.Because of the requirement of information accuracy in financial announcement,this paper proposes a method of using rules and models simultaneously to extract the elements of event sentences in financial field accurately by constructing domain dictionaries.2.Based on the extracted triple information,the domain knowledge base is constructed and a complete construction system is proposed.This paper builds a domain knowledge base based on the data extracted from announcement text and some original data.3.Based on domain knowledge base,this paper proposes a solution of domain question answering system.Question answering system based on knowledge base is a practical application of knowledge base.According to the characteristics of domain and the needs of users,this paper presents a solution of domain knowledge base question answering system.Generally speaking,this paper proposes a rule method based on professional dictionary,and realizes information extraction in the financial field with in-depth learning model,and constructs a knowledge map in the financial field.Based on this knowledge map,this paper designs and implements a question-and-answer module.The test section of the system verifies the effectiveness of the algorithm proposed in this paper,and meets the needs of the system,which has strong practicability.