Predicting corporate financial distress based on rough sets and wavelet support vector machine
Jian-Guo Zhou, Jiming Tian · 2007
This paper puts forwards a classifier hybridizing rough sets (RSs) and wavelet support vector machine (WSVM). Rough sets method is used as a preprocessor to select the subset of input variables. Then a method that generates wavelet kernel function of the SVM is proposed based on the theory of wavelet frame and the condition of the SVM kernel function. The Mexican Hat wavelet is selected to construct the SVM kernel function and form the wavelet support vector machine (WSVM). The effectiveness of the model is verified by experiments through the contrast of the results of SVMs with different kernel functions and other models.