一個有效的財務報表舞弊偵測模型

文化大學機構典藏 CCUR > 商學院 > 會計學系暨研究所 > 博碩士論文 > Item 987654321/48504

Please use this identifier to cite or link to this item: https://irlib.pccu.edu.tw/handle/987654321/48504

Title:	一個有效的財務報表舞弊偵測模型 An Effective Model of Financial Statements Fraud Detection
Authors:	李柏辰
Contributors:	會計學系
Keywords:	財務報表舞弊類神經網路決策樹支援向量機貝氏信度網路
Date:	2020
Issue Date:	2020-08-26 14:44:27 (UTC+8)
Abstract:	財務報表為企業提供給使用人了解企業之經營狀況，有些企業為隱匿其財務情形，選擇偽造其財務報表。因此，有學者針對這一方面進行研究，期望能對財務報表舞弊進行偵測達到預防的效果。本研究資料選用臺灣經濟新報資料庫(Taiwan Economic Journal, TEJ)，並依據投資人保護中心公告發生財務報導不實企業做為研究樣本，選用期間及對象為其公告資料有記錄起1998年至2018年，總計21年之臺灣上市、上櫃之全部產業。近年資料探勘(data mining)在研究上作為預測方式有顯著的成效，故本研究以類神經網路(artificial neural network)作為第一階段的重要變數篩選，再以決策樹CART、支援向量機(support vector machine)以及貝氏信度網路(bayesian belief network)來構建第二階段之預測模型。實證結果顯示，由類神經網路搭配決策樹CART(ANN-CART)擁有最佳整體預測能力之模型，準確率為95.45%。 Financial statements are provided for users to understand the operational situation. Some companies have chosen to falsify their financial statements in order to hide their financial situation. Therefore, some scholars have researched in this aspect, hoping to detect fraud in financial statements and prevent it. The selection data of this study comes from the Taiwan Economic Journal (TEJ), and based on the financial report of the investor protection center. The period of selection is from Taiwan whole industry of the upper cabinet in 1998 to 2018, a total of 21 years. In recent years, data mining has achieved remarkable results in research as a prediction method. Therefore, this study uses the artificial neural network as the important variable screening in the first stage, and then using the decision tree CART, support vector machine and bayesian belief network to build the second stage prediction model. The empirical results show that the model of artificial neural network matching decision tree CART is the best prediction model with an accuracy rate of 95.45%.
Appears in Collections:	[會計學系暨研究所 ] 博碩士論文

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	517	View/Open

View Licence

Loading...