The Artificial neural network research of audit fee and non-audit fee for listed companies

• Main Authors: Heng-Yi Liao, 廖恒毅
• Other Authors: Shi-Sheng Chen
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/97754243170104548109

碩士 === 國立勤益科技大學 === 研發科技與資訊管理研究所 === 99 === Good model of information disclosure helps to reduce information asymmetry, increase soundness of capital market, improve information transparency, and also enhance the monitoring of enterprise. The accountant who notarizes the financial report plays an important role in the information which exposed from enterprise. However, the investors might be question about audit information, have doubt about financial reports, and discredit dependability of accounting/auditing standard rule from public company. Exposing information of audit fee and non-audit fee disclosure, not only increase reputation of enterprise, helps investor and stockholder to evaluate risk analysis and value of enterprise, but also improve the result of evaluation of information disclosure. To see if the professional work from accountant or agency firm is equitability and independence, the disclosure of audit fee provides investors to assess it. Researchers all study regression analysis in the past, the weakness of it is that the structure information is complicated and mistaken. Artificial Neural Network (ANN) is efficiently used on financial analysis in recent year. Most of the studies make accuracy of prediction as the merits of regression analysis and artificial neural network. This research is study public companies in electronics industry from 2005 to 2009 about the relationship between Logistic regression and Back- Propagation network (BPN). This research study predictability and explanation from discuss audit fee and non-audit fee and revealing of audit fee. The evidence shows accuracy of prediction from ANN with less major index better explain model of revealing audit fee than regression analysis. ANN model can explain revealing audit fee with same accurate rate as regression analysis from four important variables.