Feature selection and optimization method based on CWT
Junyu Fu, Xiaonan Xu, Yuechao Chen · 2017
Wavelet transform is widely used in feature extraction and pattern recognition, the choice of wavelet basis directly affects the processing result. This paper constructs feature vector based on the continuous wavelet transform coefficients, The method of selecting the best wavelet base is extracted combined the general characteristics with distance criterion and proposed two feature optimization methods. Three types of actual underwater acoustic targets signal is used to validate the optimization method. The classification results show that the two methods can improve the recognition accuracy under the optimal wavelet basis. For different characteristics, the optimization effect is different.