基于PCR的Madaline网络算法及其在分解严重重叠光谱峰中的应用  

Madaline Network with PCR and Separating the Serious Over lapped Spectra

作　　者：方慧生[1] 吴玉田[1] 黄春明[1] 

机构地区：[1]第二军医大学药学院药物分析室,上海200433

出　　处：《计算机与应用化学》2000年第1期96-96,共1页Computers and Applied Chemistry

摘　　要：当分解重叠严重的光谱时 ,经典的Madaline网络算法需很长的训练时间 ,而且误差较大 ,引进PCR技术则不仅能减少训练时间并且能够提高精度。This paper developed a new technique i e. Madaline network with principle component analysis (PCA) to solve the problem of separating the serious over lapped spectra. The analytical results show that while the over lapped spectra is not serious, the performance of this present method is as good as the Madaline network, but, if the spectra is seriously overlapped, it's performance will be significant better than classical method.

关 键 词：Madaline网络算法 重叠光谱峰 光谱分离 PCR技术 

分 类 号：O433.4[机械工程—光学工程]