Extreme Learning Machine for financial distress prediction for listed company
Ganglong Duan, Zhiwen Huang, Jianren Wang · 2010
To overcome the shortages of the existing financial prediction models such as strict hypothesis, poor generalization ability, low prediction accuracy and low learning rate etc., a new early warning model of financial crisis have established for listed company using Extreme Learning Machine. From five dimensions of solvency, operating-ability, profitability, cash-ability and grow-ability, fifteen financial indexes were selected as the input variables; and the output variable was defined as whether the listed company had been special treated or not. The empirical analysis results show the training and validation accuracy of the model are 100% and 92% respectively, which concludes that learning and generalization abilities of this model are excellent, and which can meet the requirements of financial distress prediction for listed company.