基于改进的BRNN网络的二级结构预测

Predicting protein secondary structure based on improved BRNN

针对双向反馈神经网络(BRNN)的结构复杂、收敛速度慢的特点,本文提出了一种改进的BRNN网络。将BRNN左、右子网络的隐层删除,直接将输入连接到状态层,并且采用BP改进算法中的弹性算法进行训练。以90条序列共15377个氨基酸进行交叉验证。仿真结果表明,改进网络以及采用的弹性算法可以有效地缩短网络收敛时间。

Because bidirectional recurrent neural network(BRNN) has complicated structure and slow convergence,this paper gave an improved BRNN structure by deleting it's right and left hidden-layer and took the RPROP algorithm to train the network.The data of 90 sequences of 15377 amino acids were used to test the method.The experiment results showed that the improved network and the RPROP method could highly increase the convergent speed.