摘要
针对神经网络隐含层确定的经验性,提出一种基于粗决策树的隐含层确定方法。利用粗糙集与决策树相结合提取出的规则来确定隐含层神经元的个数,为神经网络隐含层确定提供一种方法。并用汽车评估的实例进行验证,实验表明,这种确定方法是有效性。
Aiming at the experience of the hidden layer of neural network, a method for determining the hidden layer based on rough decision tree is proposed. The number of neurons in the hidden layer is determined by using the rules extracted by the combination of rough set and decision tree to provide a method for neural network hidden layer determination. And verified with examples of vehicle evaluation, experiments show that this method is effective.
出处
《数码设计》
2016年第3期29-34,共6页
Peak Data Science
基金
河北省自然科学基金资助项目(F2014209086)
关键词
神经网络
隐含层
规则
决策树
粗糙集
neural network
hidden layer
rule
decision tree
rough set
