摘要
针对新兴技术企业信用风险评估的必要性和现有评估方法仅局限于财务指标且指标之间高度相关的缺点,借鉴了可变精度粗糙集(VPRS)模型具有噪声数据的强适应能力和强抗干扰能力的优点,提出了一类基于VPRS的新兴技术企业信用风险识别方法,并用已上市的部分新兴技术企业对其进行实证检验,检验结果表明了该方法具有较好的识别能力。该方法首先运用VPRS理论的最新研究成果,并借助于粗糙集分析软件ROSETTA,对由训练样本组成的数据关系表进行数据补缺、离散化及属性的β约简等处理,从而导出识别规则,形成识别规则库;然后集成二叉树构建一类新兴技术企业信用风险识别方法;最后用测试样本对方法的识别精度进行检验。
Considering of the necessities of the emerging technology firms' credit fisk's assessment and the shortcoming of existing assessment approaches which mostly base on the financial assessment indexes and the assessment indexes are highly correlative, this paper introduces an identification method of emerging technology firms based on variable precision rough sets (VPRS), which uses the advantage of VPRS in dealing with noise data, then the empirical testifies the method by the emerging technology listing firms, the result demonstrates the approach identifying credit risk well. The approach applies the fruit of VPRS theory firstly, with the analytical software ROSETTA, and then completes, discretizes, attributes reduces and rules generates the data relation table which is composed of training sample and identifying indexes to deduce the rules identifying the credit risk of the emerging technology firms ; then rules house is built, constructs the identification approach integrating the binary tree secondly, finally testifies the integrated approach by using the testing sample.
出处
《管理工程学报》
CSSCI
北大核心
2010年第1期70-76,共7页
Journal of Industrial Engineering and Engineering Management
基金
国家自然科学基金资助项目(70671017)
成都市科技计划资助项目(cdkj-07-03/04)
关键词
信用风险识别
可变精度粗糙集
属性约简
规则提取
二叉树
credit risk identification
variable precision rough sets
attributes reduction
rules acquisition
binary tree