摘要
分类是数据挖掘中重要的课题,为协调决策分类,提出了一种基于粗糙集理论和BP神经网络的数据挖掘的方法。在此方法中首先用粗糙集约简决策表中的冗余属性,然后用BP神经网络进行噪声过滤,最后由粗糙集从约简的决策表中产生规则集。此方法不仅避免了从训练神经网络中提取规则的复杂性,而且有效的提高了分类的精确度。
Classification is an important theme in the data mining. In order to cooperate the decision classification, an approach to data mining which integrates rough set and BP neural network is proposed in this paper. Different from previous work, in this method, firstly irrelevant and redundant attributes are removed from the decision table by rough set. Secondly, a BP neural network is used to delete noisy attributes in the decision table. Finally classification rules are generated from the reduced decision table by rough sets.
出处
《计算机与数字工程》
2009年第10期88-90,共3页
Computer & Digital Engineering
关键词
BP神经网络
粗糙集
分类
数据挖掘
决策表
BP neural network
rough set
classification
data mining
decision table
