基于SVM的财务报告舞弊识别模型的构建与应用-会计学专业论文.docxVIP

基于 SVM 的财务报告舞弊识别模型的构建与应用 摘要 本文选取 1994～ 2013 年度受到中国证监会、上海证券交易所、深圳证券交易所和 财政部公开批评、公开谴责、行政处罚等的上市公司共 96 家，采用 1：1 的比例选取 96 家配对公司进行分析，本文通过运用舞弊 GONE 理论，分别分析了贪婪、机会、需要和 暴露这四个因素与财务报告舞弊的关系，分别设定指标，并利用支持向量机的方式构建 模型对其进行分析识别，然后与 Logistic 构建的模型的识别率进行比较。其中事前预警 模型采用基于 GONE 理论的贪婪、机会、需要和暴露这四个因子的替代变量进行构建； 事后侦查模型通过财务报告中具有显著性差异的财务指标进行构建。 识别后得出如下结论: （1）采用支持向量机的方法构建事前预警模型和事后侦查模型最终得到的判别准 确率分别为 73.26%、71.43%； （2）采用 Logistic 方法构建事前预警模型和事后侦查模型最终得到的判别准确率 分别为 76.6%、66.1%； （3）支持向量机构建的两个模型的第一类误判率都低于第二类误判率，说明本文 采用支持向量机所建立的模型识别效果比较好； （4）Logistic 方法构建的两个模型的第一类误判率都低于第二类误判率，说明本文 采用 Logistic 方法所建立的模型识别效果也比较好； （5）采用支持向量机方法建立的模型识别效果要好于 Logistic 方法构建的模型的 识别效果。 关键词：GONE 理论，财务报告舞弊，支持向量机，舞弊识别模型 I Financial reporting fraud research based on the method of the SVM ABSTRACT This article selects from 1994 ~ 2013, which criticized by public、censure by public and administrative penalties by the China Securities Regulatory Commission 、 Shanghai Stock Exchange, Shenzhen Stock Exchange and Ministry of Finance. A total of 96 fraud company select , with the proportion of 1:1 matched company 96 selected to analysis. This paper bases on the theory of fraud GONE, we analyze the relationship between of the greed 、 chance、needs and exposure and financial reporting fraud. We set index respectively, and use way of support vector machine (SVM) to construct the model of financial reporting fraud and non financial reporting fraud company. And we draw the conclusions: (1) when the companys managers are risk preference , the managers own significant stakes, and the managers join together of two position, the managers will abuse power for personal gains, try to use the right to meet the interests of the individual, purse the maximization individual utility, which means the managers own higher level of greed, the greater the chance of fraud. (2) The chairman change or not, related transactions more, The greater effect of the related transaction, than the greater the possibility of financial reporting fraud. The less the times of shareh