Research On XBRL Document Automation Data Framework Based On Semantic Web Technology

The rapid informationization of the financial industry has promoted the research step of including XBRL documents into the open data system,and the related research has gradually shifted from theoretical exploration to scene application.However,the expansion of data scale and semantic complexity brought about by the era of big data put forward two requirements for real-time extraction and mining of XBRL documents: one is the requirement that XBRL documents have higher semantics,and the other is a set of framework capable of large-scale processing of XBRL instance documents.The conflict between these two requirements poses a great challenge to the semantics of XBRL documents.First of all,the semantics of XBRL documents are generally achieved by constructing a semantic model and adding rich concepts and attributes to the semantic model.However,the increase of the data scale leads to problems in the read-write,update,storage and query of a complete semantic model,that is,the current application scenarios are only used in small-scale data sets.Second,frameworks that deal with XBRL instance documents on a large scale focus more on solving distributed parsing of XBRL documents using computational models,ignoring the rich semantics and complex relationships defined by XBRL documents.Finally,the conflict between scaling and semantic quality makes XBRL documents that cannot be linked to external data become islands of data.In this paper,the semantic network technology is firstly proposed to combine XBRL document semantics with large-scale data processing framework,the basic data structure of XBRL document is analyzed,and the basic data model of XBRL document is designed.On this basis,a complete semantic model is constructed for XBRL document based on OWL/RDF syntax.In addition,rich concepts and relationships are added to the XBRL standard classification ontology to realize the semantic modeling of XBRL documents.Secondly,this study proposes a data framework for automated semantic large-scale XBRL instance documents.Through automatic entity recognition and entity optimization algorithm based on external knowledge base and inline frequency,the transformation from XBRL instance documents to RDF data can be achieved without manual intervention.And design the graphical database storage mode of XBRL document and publish its link on the Web,liberating the huge data mining potential of XBRL data.Finally,an experiment was designed to evaluate the processing efficiency and data conversion quality of the data framework by verifying the query accuracy of converted RDF and the impact of different conversion mechanisms on RDF data quality,which proved the effectiveness and efficiency of the automated semantic data processing framework proposed in this study...