人工神经网络模型在非ABO_3型希土复氧化物电阻率特性分析中的应用

RECOGNITION OF THE RESISTANCE RATE OF R.E.COMPLEX OXIDES OF NON-ABO_3 TYPE BY NEURAL NETWORK

目前,人工神经网络在模式识别及其他领域的应用引起了普遍的关注.本文叙述“反向传播”神经网络模型,算法和在非ABO_3型希土复氧化物电阻率分析中的应用,实验结果表明,神经网络的判别正确率为100％,因此,该方法可望成为化合物分析,材料设计的辅助手段.

At present, neural network has brought wide attention in pattern recognition and other fields. The back-propagation neural network model, algorithm and its application to the predictin of the resistance rate of R.E. complex oxides of non-ABO3 type were presented in this paper . The experimental results showed that the correct rate of neural network reached 100%. Therefore the neural network method might be referred as an assistant technique for analysis of compound and material design.