蛋白质二级结构预测的人工神经网络方法研究

Study of Neural Network Methods in Predicting Protein Secondary Structures

本文比较了五种神经网络方法预测蛋白质二级结构的准确率,并做出初步评价。五种神经网络分别是:误差反传前向网络(BP),径向基函数网络(RBF),广义回归神经网络(GRNN),串并联叠层网络(CF),Elman网络(ELM)。结果显示:GRNN的预测准确率达85.7%,优于其它网络。本文还讨论了训练集样本数及参数的优化对GRNN预测准确率的影响。

In this paper, five neural network models, such as back-propagation neural network（BPNN）, radial basis neural network（RBFNN）, generalized regression neural network（GRNN）, cascade forward back been than also propagation neural network（CFNN） and Elman backpropagation neural network（ELMNN）, have evaluated in predicting protein secondary structures. The prediction accuracy of GRNN is better the others. In addition, some affecting factors（the training sets and the parameters of network） are discussed.