摘要
本文提出了一种基于粗糙集理论的模式分类样本特征选择方法,该方法应用粗糙集理论和方法,对给定的学习样本进行特征选择,根据这些特征构造神经网络模型进行训练,并在网络的工作阶段,根据这些特征对待识样本进行分类。在模式分类中,该方法能够减少网络的训练时间并改善网络的泛化能力。
A sample feature selecting method for pattern classification based on rough sets is presented in this paper. Features are selected from the given training data by applying the theory and method of rough sets. Neural networks are constructed and trained according to these features, and trained networks are used to classify the other data on the basis of these features. In the pattern classification, this method can reduce the training times of neural networks and improve the generalization ability of networks.
出处
《计算机应用与软件》
CSCD
北大核心
2003年第2期9-10,38,共3页
Computer Applications and Software
基金
湖南省自然科学基金资助(编号:00JJY2059)
