摘要
提出一种改进的隐含层神经元个数的选取方法.该方法基于隐含层神经元输出之间的相关关系,估计隐含层结点数,把隐含层结点的改进算法用于物种选取的实例中,通过MATLAB编程实现,并与回归分析做出的结果进行比较.结果表明,该方法对建立BP神经网络结构有一定的借鉴作用.
An improved method for choosing the number of nodes in hidden layer was proposed. The method is based on the linear correlation of the output of hidden layers to estimate the number. The last part put the improved method into the application of species select, using MATLAB program to realize and making comparison with the result of regression analysis. The results show that the method makes advance in guiding to the structure of the BP neural network.
出处
《西安工程大学学报》
CAS
2008年第4期502-505,共4页
Journal of Xi’an Polytechnic University
基金
陕西省教育厅自然科学专项基金资助项目(06JK286)
关键词
BP神经网络
隐含层结点
线性相关
物种选取
BP neural network
the node of hidden layer
linear correlation
species select
