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

Abstract

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

Keywords:

• dynamic financial distress prediction
• sliding time window
• pharmaceutical companies
• entropy-based weighting
• support vector machine

Acknowledgement

The authors gratefully thank anonymous referees for their useful comments and editors for their work.

Conflict of interest disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research is partially supported by the National Natural Science Foundation of China [grant number 71371171], [grant number 71171179]; the Humanity and Social Science Foundation of Ministry of Education of China [grant number 13YJC630140]; the Zhejiang Provincial Natural Science Foundation [grant number LY13G010001], [grant number LR13G010001]; the Zhejiang Provincial Philosophy and Social Science Foundation – Zhijiang Young Talent of Social Science [grant number 11ZJQN082YB]; and the Zhejiang Provincial Qianjiang Talent Foundation of China.