Financial Crisis Dynamic Prediction Based on Sliding Window Technology and Mahalanobis-Taguchi System

In order to improve the prediction accuracy of current existing model, the financial crisis prediction dynamic model is proposed. By means of the data streams processing method, the sliding window technology is used for real-time updated samples in this paper, and then the optimal features of samples are selected by Mahalanobis-Taguchi System. The financial crisis prediction dynamic model is built on financial data of the Chinese listed companies. Experiment results show that dynamic selection of the samples and features can reduce the adverse effects of the concept drift in financial data streams. The prediction accuracy of the dynamic model based on sliding window technology and Mahalanobis-Taguchi System is better than the static prediction models.