یک مطالعه در استفاده از روش داده کاوی در افشای اطلاعات برای سرمایه گذاران بازار سهام تایوان

The financial literature and practices have shown the importance of corporate governance for decades, not only for firm’s management but also for investor protection. Information disclosure plays a key role in all of the governance mechanisms. With good information disclosure, the asymmetric information and the agency cost between the insider and the outsider of firms can be reduced effectively. However, the information disclosure status of listed companies is hard to be evaluated or judged by investors before the annual official announcement is reported in the next year. The main purpose of this study is to explore the hidden knowledge of information disclosure status among the listed companies in Taiwan’s stock market. In this paper, we employed decision tree-based mining techniques to explore the classification rules of information transparency levels of the listed firms in Taiwan’s stock market. Moreover, the multi-learner model constructed by boosting ensemble approach with decision tree algorithm has been applied. The numerical results show that the classification accuracy has been improved by using multi-leaner model in terms of less Type I and Type II errors. In particular, the extracted rules from the data mining approach can be developed as a computer model for the prediction or the classification of good/poor information disclosure potential and like expert systems.

In recent years, a number of financial scandals, such as Enron and WorldCom, have made the corporate governance to become a noticeable topic to both practice and academy in financial management. Most of the problems were caused by the asymmetric information between the insiders and the outsiders of the firms. The Organization for Economic Co-operation and Development (OECD) has identified the corporate governance as a direct monitoring system to maximize the value of firms through the mechanism of transparency and efficiency. OECD also declared six principles for corporate governance in 2004 (http://www.oecd.org/dataoecd/32/18/31557724.pdf). Among these principles, information disclosure plays a key role to ensure whether the internal and the external governance mechanisms work well or not. The two mechanisms defined by World Bank in 1999 (http://www.sovereignglobal.com/media/framework_for_implemenation.pdf) are as follows: internal governance includes board of director’s monitor, compensation of manager and incentive plan, etc.; external governance includes external auditing, professional investors and government, etc. Obviously, the above two governance mechanisms have to rely on correct and proper information disclosed by the firms. However, the judgment of information disclosure status is with some incentive-bias problems. For instance, firms prefer releasing good news and concealing the bad ones, such as those firms who violate the regulations or suffer the financial distress. Thus, the whole worldwide stock market authorities keep reviewing the firms’ information disclosure status and related regulations to protect the interests of investors and public. The evaluation of information disclosure is more complex than just posting their financial statements or annual reports on the publication or website. According to the Public Company Accounting Reform and Investor Protection Act of 2002, except the accurate financial statements, the factors for evaluating the information disclosure status must include the off-balance sheet transaction, pro forma financial information, major stockholder transaction, manager disclosure valuation, code of ethics of CEO, and opinions of financial experts. Obviously, it is difficult for the individuals or even professional investors to understand the full dimensions of a firm’s disclosure situation, because we never know whether the firm shows out the whole picture truly or not. Only the authority has the jurisdiction over the authenticity for a firm’s disclosure. For better understanding the transparency of listed firms, the Securities and Futures Institute (SFI) of Taiwan initiated the investigation of the information disclosure level among all the listed firms in stock market. Although the investigation report is announced annually, it always takes more than one year to have this official report. So, the investors cannot do anything by this antiquated announcement. Moreover, the information disclosure also has some significant relationship with some specified firm’s business characteristics. For example, Chen and Jaggi (2000) and Eng and Mak (2003) have shown the association among the ownership structure, board composition, firm characteristics and disclosure. In this paper, we developed a transparency evaluation approach by using data mining techniques. It is wished to help investors to identify the corporate governance situation, and then make their investment decisions correctly in time. There are rare studies on the use of this new emerging technology on this topic. Through mining the ownership structure and firm characteristics data by using decision tree-based techniques, we can classify and predict which ownership structure and characteristics will bring out which type of transparency in time, and does not have to wait for the official report until next year.

The information disclosure plays a very important role to ensure whether the internal and external governance mechanisms of the listed firms work well or not. If the relationships between the status of information disclosure and the companies can be inferred correctly and quickly, then the stock investors can have much better guideline for their investment with more profits. By using the decision tree-based rule mining approach, the significant factors with the corresponding equality/inequality and threshold values are decided simultaneously, so as to generate the decision rules. Unlike many mining approaches applying neural networks related approaches in the literature, the decision tree approach is able to provide the explicit classification rules. Moreover, a multi-learner model constructed by boosting ensemble approach with decision tree algorithm is used to enhance the accuracy rate in this work. Based on the extracted rules, a prediction model is then built to discriminate good information disclosure data from the poor information disclosure data with great precision. Moreover, the results of the experiment show that the classification model obtained by the multi-learner method has higher accuracy than those by a single decision tree model. Also, the multi-learner model has less Type I and Type II errors. It indicates that the multi-learner model is appropriate to elicit and represent experts’ decision rules, and thus it provides effective decision supports for judging the information disclosure problems in Taiwan’s stock market. By using the rule-based decision models, the investors and public can accurately evaluate the corporate governance status in time to earn more profits from their investment. It has a great meaning to the investors, because only prompt information can help investors in correct investment decisions.