The dynamic financial distress prediction method of EBW-VSTW-SVM

摘要: Financial distress prediction FDP takes important role in corporate financial risk management. Most of former researches this field tried to construct effective static SFDP models that are difficult be embedded into enterprise information systems, because they based on horizontal data-sets collected outside the modelling by defining as absolute conditions such bankruptcy or insolvency. This paper attempts propose an approach for dynamic evaluation and entropy-based weighting EBW, support vector machine SVM enterprise’s vertical sliding time window VSTW. The DFDP method is named EBW-VSTW-SVM, which keeps updating model dynamically with goes only needs historic data itself thus easier systems. EBW-VSTW-SVM consists four steps, namely relative VRFD construction training data-set according VSTW, future point. We carry out case studies two listed pharmaceutical companies experimental analysis some other simulate window. results indicated proposed was feasible efficient help managers improve