Financial Early-Warning Model Based on Q-Gaussian Kernel Fisher Discrimination

Q-Gaussian function has the extensive scope of application compared with Gaussian function. It can become many different radial basis functions when we choice the different parameters. Q-Gaussian is chose as kernel to establish the financial early warning model of listing Corporation in this paper. Through the contrast of the Fisher model based on Gaussian kernel, polynomial kernel and the linear Fisher discrimination, it is proved that the Fisher model based on Q-Gaussian kernel has higher accuracy of prediction.