摘要
采用改进型反向传播人工神经网络,以含铁矿物穆斯堡尔参数作为样本,通过神经网络的训练,能很好地识别矿物,从而有效地提高了对矿物的识别本领。
Minerals have been identified using Mssbauer parameters and artificial neural networks(ANN). The reported Mssbauer parameters of minerals were used to train an ANN called the improved back-propagation network. The excellent corrective ability indicates that it is probably an alternative method in Mssbauer data processing.
出处
《核技术》
CAS
CSCD
北大核心
2000年第7期467-474,共8页
Nuclear Techniques
基金
国家自然科学基金!19835050
国家教委博士点基金!98-00
关键词
人工神经网络
穆斯堡尔谱学
矿物识别
Artificial neural network, M■ssbauer spectroscopy, Improved back-propagation algorithm