Quality Evaluation Of XBRL General Taxonomy Elements Based On Information Entropy

XBRL(Extensible Business Reporting Language)is widely used worldwide for its strong advantages in financial information disclosure,analysis and cross-platform communication.The most important part of XBRL is the XBRL taxonomy,and the core of the taxonomy is the XBRL general taxonomy elements.When formulating the taxonomy,it is first necessary to determine which items are to be used as general elements.With the continuous updating of the XBRL taxonomy version,XBRL general taxonomy elements are constantly changing.The quality of elements represents the quality of the taxonomy.Therefore,the scientific problem to be solved in this thesis is: how to evaluate the quality of XBRL general taxonomy elements?First of all,through literature review,we found that previous related research has the disadvantages of single research perspective,inefficient research methods,and it's not systematic,thus clarifying the research ideas of this thesis;then,from the perspective of information,designed to the quality evaluation framework of XBRL general taxonomy elements,using theories of information entropy,text similarity calculation,and statistical language model,construct measurement of quality dimension.Next,an empirical study of elemental completeness and comprehensibility was conducted to measure the degree to which the XBRL general taxonomy elements meets enterprise information disclosure needs in practice and the intelligibility of the language,and element coverage study from both the financial reporting framework and financial information content,understand the use of elements;Finally,based on the research results,research suggestions are given to provide reference for XBRL general taxonomy elements and even the improvement of the whole XBRL taxonomy.The main research results of this thesis are:(1)From the perspective of information,a quality assessment framework for XBRL general taxonomy elements is designed.The quality framework consists of completeness,comprehensibility and coverage of the elements.The meaning of the quality dimension of this research is different from the previous research.Compared with the manual element matching research method,the research method of this paper is more efficient.(2)An empirical study of completeness and comprehensibility of the elements shows that: Compared with the disclosure practice of PDF financial reporting,the 2015 XBRL general taxonomy elements is insufficient in terms of the content of carrying information and not well meet the needs of enterprise information disclosure;comprehensibility also needs to be improved;the 2015 XBRL general taxonomy elements is better than the 2010 XBRL general taxonomy elements,which is mainly reflected in the increase in information content and the increase in the regularity of the language,which improves the comprehensibility of the taxonomy.(3)From the financial reporting structure framework and the financial information content,the element coverage study was conducted.It was concluded that compared with the PDF financial reporting practice,the coverage of the 2015 XBRL general taxonomy elements was less than 100%,element usage was not good,here are elements redundancy phenomenon.The phenomenon of redundancy focuses on the disclosure of smaller elements,and that the use of elements not only has industry differences and company preferences,but also has industry internal consistency.The objective reason for the mutual restriction between element completeness and correlation is verified by combining the research results of elements completeness.