基于决策树的财务危机预警模型研究

A Model of Financial Distress Early-warning Based on Decision Tree

提出一种基于改进了的决策树的财务预警模型。该方法通过正规增益标准对企业的财务指标进行排序降维,从而避免了冗余信息的影响,直接生成最小决策树,抽取预警规则。最后以C4.5算法为参照,进行了仿真试验,结果表明,本算法在预警效率和预测精度上都有明显的提高。

A new model of enterprise financial distress early -warning based on improved decision tree is presented. The method ranks financial data of enterprise based on the Normalized Gain, avoids redundancy information＇s influence, generates the least decision tree directly and extracts the warning rule. Finally comparing the methods with C4.5 algorithm, studies are empirically carried out, and results show that this methods largely improves the warning efficiency and precision.