以關鍵查核事項偵測公司信用風險──機器學習之應用

Detecting Credit Risk by Using Key Audit Matters-An Application of Machine Learning

"關鍵查核事項(Key Audit Matter, KAM)，指依會計師的專業判斷，對本期財務報表之查核最為重要的事項。揭露KAM之目的，在於對會計師所執行之查核提供更高的透明度，並增進查核報告的溝通價值。然關鍵查核事項所的揭露，是否真能有助於資訊使用者的決策，仍未有清楚的論證。本文以2016年到2018年我國的上市櫃公司為樣本，透過文字探勘方式，解析每一篇關鍵查核事項的文字資訊，適當連結KAM所傳達的公司信用風險程度。並以機器學習方法中的遺傳基因演算法優化的支援向量機模式，去瞭解關鍵查核事項是否能有效判別公司信用風險的高低。分析的結果發現，關鍵查核事項確實可以有效的區分公司信用風險的高低，對於協助投資人在制定決策上具有重要價值。這結果支持我國推動關鍵查核事項之揭露，能達成提高財務報告透明度與協助資訊使用者進行相關決策的目標。"

"Key Audit Matters (KAM) are defined as“Those matters that, in the auditor’s professional judgment, were of most significance in the audit of the financial statements of the current period.”The disclosure of key audit matters (KAM) aims to enhance information transparency and value relevance of audit reports. However, the disclosure of KAM is still inconclusive with respect to whether it can improve decision usefulness for users. Using publicly-traded companies from year 2016 to 2018 as the sample, this study first identifies credit risks via analyzing the text content of each key audit matters by text mining. Next, the Genetic Algorithm is employed to optimize Support Vector Machine, aiming to detect the company's credit risks. The results of this study indicate that the key audit matters are effective in identifying credit risks, and therefore are value-relevant in decision making for investors. In summary, results support the arguments that the disclosure of key audit matters is likely to improve financial transparency and decision making for users."