Predicting corporate financial distress by PCA-based support vector machines

This paper proposed a hybrid principle component analysis based support vector machines to predict the corporate financial distress. In the proposed approach, principle component analysis is used for feature selection to reduce the computation complexity of support vector machines and then the support vector machines is used to identify corporate financial situation based on the historical data. To evaluate the performance of PCA-based support vector machines, we compare its results with that of conventional methods and neural network models. The experimental results suggest that PCA-based support vector machine outperforms other forecasting model.