Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/18786
標題: 從公司治理觀點建構訴訟預警模型-類神經模糊之應用
Corporate Governance and the Prediction of Litigation Presence- A Neuro-Fuzzy Approach
作者: 莊家豪
Chun, Chia-Hao
關鍵字: Corporate governance;公司治理;Litigation warning;knowledge;Logistic regression;Neuro-fuzzy;訴訟預警;羅吉斯迴歸;類神經模糊
出版社: 會計學研究所
摘要: 
本研究探討上市櫃公司之公司治理機制能否有效發揮自我監督之防弊功能,藉以提供會計師做為查核策略判斷之依據,以證期局及投保中心公佈之62個訴訟案件作為起訴公司樣本,另行配對124家非起訴公司,針對起訴公司在公司治理因素之缺失進行探討,並嘗試運用羅吉斯迴歸及類神經模糊技術建構訴訟預警模型,以獲取公司治理機制缺失與發生起訴與否之關聯性。實證發現董監持股、董監規模與起訴與否呈顯著負相關,而法人股東、次大股東與起訴與否呈顯著結果,但正、負相關互見。對預測模式之發射警訊而言,羅吉斯迴歸在最早期即能提供強烈警訊,有效預示情危公司,但隨著結構性改變,如公司治理法規愈趨嚴謹,則無從端倪該細微變化;類神經模糊可透過學習訓練過程,以在最逼近起訴時點提供最強烈警訊,告知利害關係人及早做好防範措施,減低經濟損失及資源耗費。類神經模糊亦提供公司治理機制缺失之內在規則及重要變數關係之3D圖,協助會計師評估客戶風險以提升查核效果及效率,及作為未來立法之參考。

This study examines if corporate governance mechanisms of publicly listed companies may play the role of self-supervision, hence provide auditors with judgmental assistances for decision-making. A sample including 62 sued cases selected from Securities and Futures Bureau as well as Investors Protcetion Center and 124 non-sued companies chosen as matched-pair samples having equivalently demographic characteristics of size and industry is used to analyze the characteristics and weaknesses of contemporary corporate governance. The study applies a logistic regression and a neuro-fuzzy technique to construct litigation-presence warning models, subsequently to capture the relationship between corporate governance and litigation presence. Empirical results show that litigation presence significantly has negative relation with both the shareholding and the number of directors and supervisors. However, the relationship between institutional/secondary shareholders and litigation presence remains unclear. Further, concerning the prediction ability, the logistic regression can provide the earliest warnings in comparison with neuro-fuzzy, but such an ability would be violated if structural changes occur, for instance, law regulations become rigorous. On the contrary, the neuro-fuzzy with its unique ability of learning offers better warning while the time is getting closed to litigation occurrence. Hence benefits to the related parties could be derived from avoiding economic losses and resource wastes. In addition, the knowledge base rules and 3D plots among the variables obtained from the neuro-fuzzy also offer a promotion of auditing effectiveness and efficiency, and a guidance for regulation establishments.
URI: http://hdl.handle.net/11455/18786
Appears in Collections:會計學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.