THE RECOGNITION OF UV SPECTRA BY USING WA VELET NEURAL-NETWORK

The wavelet neural network employed in signal recognition is presented and applied to recognition of the UV spectra of tyrosine; 3,4-dihydro xyphenylalanine and trytophane. In the network, Morlet mother wavelet and line search conjugate gradient optimization method are used, The c onvergence speed is fast and the error of learning sets drops to 0. 00 0 073 only through 8 iterations, The error of testing patterns is 0. 0 00 64. The recognition results (shown in Table 1) show that the sets o f spectra can be well recognized by using the wavelet neural network.