Research On The Construction Of Personal Relationship Graphs In The Financial Domain Based On Deep Learning

In the information age,listed companies and financial regulators will publish a variety of massive announcements containing unstructured personnel resumes on the Internet.Describeing personnel attributes,entities,and events in the unstructured personnel resume by a structured form,discovering the hidden personnel relationships and building personal relationship graphs in the financial domain can mine the rich knowledge contained in financial announcements,identify potential customers,warn potential risks,and improve the efficiency of the financial industry.At present,the personnel resume information extraction tools which have certain limitations are all processed for structured and semistructured resume files,and it is hard to extract personnel attributes and events for unstructured personnel resume texts in financial announcements.In addition,the traditional relationship extraction is to discover the relationship between entities in one sentence.Most of the attributes of the relationship are not further extracted,and the personnel relationship between different financial announcements cannot be found.To solve the special problems in the information extraction of personnel resume in financial announcements,this paper is the first time to extract unstructured personnel resume into a structured personnel information template,and proposes a method for constructing personal relationship graphs in the financial domain.By fine-tuning the BERT model,the personnel attribute entities in the unstructured resume are extracted.The finetuned BERT model is used to obtain event instance vectors,accurately classify the event instance vectors,and make relations among personnel attributes by filling the hierarchical personnel information templates,and further through the filled personnel information templates to extract personnel relationships and construct personal relationship graphs.To verify the effectiveness of the method for constructing personal relationship graphs in the financial domain,an experiment was carried out by constructing a manually labeled data set.The experiment showed that the proposed method can effectively solve the problem of extracting information from unstructured financial personnel resume,and effectively construct the personal relationship graphs in the financial domain.