Mining XBRL-based Data Hierarchy

XBRL is a kind of extensible business reporting language which based onxml,it has been widely used in the financial system. The language has three layerstructure of technical specification, classification standards and the instancedocuments.Technical specification specifies the syntax and related technical standardof the XBRL, Classification standard which made up of schema file and link librarydependent on technical specification and accounting rules, Instance documents arereport language which based on XBRL technical standard and classification standard,they Used to store details of enterprise financial data and deliver information on theinternet. The fusion of data mining and XBRL hierarchy brought convenience to ourdata analysis, more and more enterprise financial data brought us to use the method ofdata mining to dig out the important information we need.The core idea of XBRL technology is first to extract the data source, andtransfom the data source document again into an XML document, then converter intoXBRL format through the document. XBRL document can be stored in the userdatabase system or Uploade to the browser for user download. Data mining isextracted from XBRL document information analysis to extract the data we need. Thecommon process of Data mining is data collection, data preprocessing, data mining,and data shows, Through Apriori algorithm of association rules in data miningcombined with XBRL hierarchical structure, this paper proposes a data mining modelbased on XBRL hierarchy architecture, which include four modules of data extractionand conversion, X-Hive data storage, association rules mining, and data shows. Thismodel integrated the XBRL hierarchy, which in accordance with all relevant datamining process.And use the method of association rules and XQuery queries for deep mining on XBRL data which storage in the X-Hive database.In the process of datamining on XBRL, To improve the algorithm of Apriori, this paper put forward aDC-Apriori algorithm based on X-Hive,which make data mining on XBRL moreefficient.The experiment show that use DC-Apriori algorithm in X-Hive to excavateXBRL data is Feasible and effective, And the efficiency of the DC-Apriori algorithmis higher than the Apriori algorithm in the relational database application.