摘要
针对传统BP神经网络收敛速度慢且容易陷入局部极小值的缺点,通过实验对其进行分析并提出可以提高算法快速性和有效性的改进方法。介绍了BP神经网络的工作原理、基本流程、算法应用领域和在该领域下的优缺点,以及改进方法的具体步骤与改进后的优势。最后通过仿真实验验证了改进算法的有效性。
For the disadvantages that traditional BP neural network is slow convergence speed and easily trapped in local minima shortcomings,the article analyzes it through the experiment and puts forward improved method that can improve the quickness and the effectiveness of algorithm.This paper systematically introduces the working principle and the basic process of BP neural network,application field of the algorithm and the advantages and disadvantages under it,finally validates the effectiveness of the improved algorithm through the simulation experiment.
出处
《软件导刊》
2016年第5期38-40,共3页
Software Guide
关键词
神经网络
BP神经网络
BP算法
Neural Network
BP Algorithm
Classification
Prediction
