属性识别预警系统(AFS)——一种新的信用评价方法

在属性集和属性测度函数的基础上建立了属性测度预警系统.利用我国深、沪两交易所2000年公布的106家上市公司作为训练样本,计算出“差”企业和“好”企业的属性测度均值,构造出属性测度函数.进而对2001年公布的13家预亏公司进行预警分析,用置信度准则及最大属性测度准则判别,预警准确率分别达到92.31%和100%.因此,属性识别预警系统在信用评价方面具有良好的应用前景.

Attribute Forecasting System(AFS)-A New Credit Scoring Method

Based on attribute sets and attribute measure function,this paper established attribute forecasting system.Using 106 on-the-stockmarket companies which are published by shenzhen and shanghai stock exchanges in 2000 as training sample,this paper counted out the mean value of attribute measure for "bad" factories and "good" factories,and constructed attribute measure function,moreover,when analysis 13 companies which forecasted loss incurred,with confidence criterion and maximum attribute measure criterion,the accuracy rate of forecasting reached 92.31% and 100%.So.AFS has quite good application prospect.