利用人工智能技術預測企業繼續經營疑慮

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 987654321/51238

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: https://irlib.pccu.edu.tw/handle/987654321/51238

题名:	利用人工智能技術預測企業繼續經營疑慮 Prediction of Enterprises' Going Concern by Using Artificial Intelligence Techniques
作者:	賴冠銘
贡献者:	會計學系
关键词:	深度學習繼續經營疑慮極限學習機長短期記憶模型線性支持向量機 Deep Learning Going Concern Extreme Learning Machine Long Short-term Memory Model Linear Support Vector Machine
日期:	2022
上传时间:	2023-03-01 14:33:49 (UTC+8)
摘要:	隨著大環境的改變，公司的是否能夠繼續經營也成為投資人最為關注的議題之一，所以會計師是否能夠出具有效的相關財務報表意見是相當重要的。本研究以2008年至2021曾經被出具繼續經營疑慮意見之上市櫃公司為主要研究對象，先使用隨機森林(Random Forest)與極限梯度提升(eXtreme Gradient Boosting)來進行重要變數的篩選，再以極限學習機(Extreme Learning Machine)、長短期記憶模型(Long Short-term Memory model)與線性支持向量機(Linear Support Vector Machine)分別來建立預測模型。實證結果顯示，隨機森林搭配線性支持向量機(RF-LSVM)有最佳的偵測能力，具有98.84%的準確度、極低之型二錯誤1.22%以及高達94.34%的F1分數。 With the change of the global environment, whether the company can continue to operate has become one of the most concerned issues for investors, so it is important whether the accountants can issue valid financial statement opinions. This research focuses on listed companies that have been issued with doubts about continuing operations from 2008 to 2021. First, Random Forest and eXtreme Gradient Boosting are used to screen important variables. A learning machine (Extreme Learning Machine), a long short-term memory model (Long Short-term Memory model) and a linear support vector machine (Linear Support Vector Machine) are used to establish prediction models respectively. Empirical results show that random forest combined with linear support vector machine model (RF-LSVM) has the best detection ability, with an accuracy of 98.84%, a very low type 2 error of 1.22%, and a high F1 score of 94.34%
显示于类别:	[會計學系暨研究所 ] 博碩士論文

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	201	检视/开启

检视Licence

在CCUR中所有的数据项都受到原著作权保护.

数据加载中.....