BP 神经网络学习算法的研究.doc

BP神经网络学习算法的研究
刘彩红资助项目:西安工业大学北方信息工程学院院长科研基金项目(BXXJJ-1112)
作者简介:刘彩红(1980-),女,陕西人,硕士研究生,研究方向为人工神经网络及其应用。Email: rainbow_dd@手机:1360 925 0662;
(西安工业大学北方信息工程学院,西安)
摘要:(目的)本文针对BP算法收敛速度慢的问题,提出一种改进的BP算法。(方法)该算法结合生物神经元学忆形成的特点,针对特定的训练样本,只激发网络中的部分神经元以产生相应的输出,而未被激发的神经元产生的输出则与目标输出相差较大,那么我们只需要对未被激发的神经元权值阈值进行调整。利用距离来度量训练样本与神经元权值之间的关系,可以找到未被激发的神经元。所以本论文提出的算法是对局部神经元权值阈值的调整,(结果)通过实验表明该算法有助于加快网络的学习速度。
关键词:BP神经网络,学习算法,距离,权值阈值调整
The Study of Algorithm of BP work
LIU Cai-hong
(Xi’an Technological University North Institute of Information Engineering ,Xi’an China)
Abstract:The paper proposed an improved BP algorithm. According to the characteristics of biological neuron in learning and memory formation, only some neurons were stimulated to produce the output for the specific training samples, while the other part of the neurons weren’t stimulated. There are large difference between this part of the neuron’s output and target, and then we need this part neurons weight and threshold value to adjust. Use to measure the distance from the training sample and the relationship between the weights of each neuron, can find not stimulated the algorithm proposed in this paper only adjust the weight and the threshold value of the local neurons, and this can accelerate the learning speed of work.
Keywords: BP work, Learning Algorithm, , Distance, Weight and Threshold Adjustment
1 引言