Research On Knowledge Graph Construction Of Listed Companies And Question Answering Technology

In the information age,information technology develops rapidly.Fintech has also been greatly developed.The traditional way of obtaining information through search engine retrieval brings users a huge number of results with complex content.Integrating knowledge with knowledge graph and using intelligent question-answering system for question-answering interaction can help users quickly obtain effective information and improve work efficiency.With the popularization and improvement of various online channels,the customer service business model has shifted from a purely human service model to a combination of automated intelligent customer service and manual customer service.Constructing a question-answering system to help users accurately obtain information and effectively solve users’ questions has become a hot research issue.The main research contents of this thesis are summarized as follows:(1)A knowledge graph of listed companies is designed and constructed.The thesis uses crawlers to crawl the basic information of 4187 listed companies,basic staff information,public company announcements within six months and records of punishments received.Use named entity recognition,relationship extraction and other methods to perform data processing and obtain announcement content information.According to the designed data pattern of the knowledge graph of listed companies,the Neo4j is used to construct the information knowledge graph of listed companies.(2)A question answering system based on knowledge graph and a question answering system based on FAQ are researched and implemented.The thesis uses the Bi-LSTM+CRF model to identify the named entity of the announcement text and extract the information content of the announcement.The question-and-answer module is implemented using methods such as rule-based methods,pipeline-based methods,and short text similarity matching.The thesis proves the effectiveness through experiments.(3)An intelligent question-and-answer system based on the knowledge graph of listed companies and FAQ data is designed and implemented.The thesis interacts with users in the form of WeChat official account.A feedback mechanism has been set up to record user feedback in order to maintain and improve the system.