Combining Unascertained Measure and Neural Network for Prediction of Corporate Strategic Risk

This paper establishes a framework of nonlinear combination forecasts that predicts corporate strategic risk based on unascertained measure and neural network. Firstly, a multidimensional indicator system is proposed for the prediction of corporate strategic risk, in which strategic risk levels is divided into strategic financial risk, competitive risk and relative performance risk. Secondly, the unascertained measure model is proposed to assess the strategic risk on the basis of three-layer indicator system. Thirdly, the multiple regression model and Logit model are introduced to assess the financial risk. Finally, the combined forecasting model based on neural network is presented to predict the strategic risk. Computational experiment result shows that the proposed method performs better than others in forecasting strategic risk for listed firms