可延伸商業報導語言(eXtensible Business Reporting Language, XBRL)最重要的應用在於企業之財務報告書。我國行政院金融監督管理委員會在資本市場全球化的趨勢下,推動我國企業於2010年正式邁入XBRL財務報告申報元年。企業申報之XBRL四大財務報表及附註揭露皆以全球統一通行的標準格式為大眾所用,將使得更多財報使用者提高對企業資訊的關注,形成一個全民稽核的時代。此時,若企業無法掌握XBRL財報品質,除了讓舞弊行為有機可趁,更可能在風險發生時引發企業重大損失。 因此本研究旨在提出一個結合XBRL與資料探勘技術之財報檢核電腦稽核系統,除了能驗證財報上的數字資料,並運用資料探勘技術對文字性的附註揭露進行分析,藉以提高二者資料間判讀攸關能力與財報閱讀性,並透過這樣的功能設計讓使用者能夠易於進一步的查核與分析。依據本研究之實驗結果證明,本研究機制是可行、有效的並能提升稽核效率,冀望透過本研究機制能夠協助內部稽核人員提早或及時發現公司財報是否有重大錯誤與舞弊之可能,並能確實提高稽核效率與降低稽核時間與人力成本。 關鍵字:可延伸商業報導語言、電腦稽核、本體論、文字探勘、財務報表
The most important application of eXtensible Business Reporting Language (XBRL) is in financial statements of business. In the trend of globalization of capital markets, Taiwan’s Financial Supervisory Commission (FSC) promotes our businesses to use XBRL to report financial statements in 2010. The four financial statements and notes disclosure which enterprise reported are within a globally harmonized common standard format for use by the public, it will make the users of financial statements to improve more concern of business information; however, the era of everyone could be an auditor is coming. At this point, if the company is unable to get a good hold of the quality of XBRL finicail statements, it will not only increase the opportunities of fraud, but also may lead to significant losses while risks happened. Thus, the purpose of this study is to propose a combination of XBRL and data mining techniques to establish a computer auditing system to inspect XBRL financial statements, this mechanism not only able to verify numeric data but also uses text mining techniques to inspect notes disclosure which belong to text data type,which is in order to improve relevance of interpretation and statements readability between these two kinds of information, and let users could easier to further audit and analysis via the mechanism functions. Based on the experimental results of this study prove that the mechanism of this study is feasible, effective and can improve audit efficiency, hoping that through this mechanism could help internal auditors to early or timely detect if there is any possibility of serious errors and fraud, and help to improve the efficiency of audit and reduce audit time and labor costs. Keyword: XBRL, Computer Auditing, Ontology, Text Mining, Financial Statments