Having been encountered the difficulty of properly predicting the cutting chatter trend at present, a novel cutting chatter trend prediction method based on support vector regression algorithm (SVR) was put forward, meanwhile, the principle about how to select the SVR parameters was discussed. Ways about how to extract features from cutting signal process based on wavelet package decomposition were studied. Gabor wavelet basis was used to decompose the cutting signal to 3-order. The energy in each spectrum section was calculated and scaled. The resultswere input to the SVR model. Sampling and ...