改进型动态隧道神经网络在蛋白质预测的应用

Application of improved dynamic tunneling neural network in protein secondary structure prediction

神经网络具有容易陷入局部极小的缺点,动态隧道神经网络通过"钻隧道"方式,让目标函数跳出局部最小,找到更小的可行域,从而避免神经网络陷入局部极小。传统的动态隧道技术隧道方向单一并且随意,因此具有不稳定性。为了有效提高动态隧道的搜索效率,提出了一种改进型动态隧道神经网络算法。该算法增加搜索的隧道数,引入夹角弹性系数控制隧道方向,考察隧道之间的相互影响。在对alpha、beta和coil型蛋白质的二级结构预测的实验中,改进型动态隧道神经网络算法预测的效果优于神经网络算法和传统的动态隧道神经网络算法。

Neural network suffers from a defect of easily immersing in local trap.Dynamic tunneling helps neural network eliminate the local trap by＂tunneling＂and jumping into a lower valley of object function.However,the traditional dynamic tunneling technique tries to search in a random and single direction thus it is instable.In order to promote the searching efficiency,an improved dynamic tunneling neural network algorithm has been proposed to enhance the instability by increasing the directions of tunneling and controlling the interaction between trajectories of the tunneling system with an angle spring coefficient.Experimental results of the prediction of alpha,beta and coil protein secondary structure show that the improved algorithm outperforms the neural network and the traditional dynamic tunneling neural network.