摘要
为了在知识约简中能够客观地反映决策规则的决策能力,提高约简的效率和识别率,针对动态知识获取这一问题,提出了一种基于决策熵的增量式知识获取算法。该方法利用决策熵能够客观地衡量决策表的决策能力的特点,在现有规则集基础上实现知识的动态更新,避免了重复计算从而提高了知识获取的识别率和效率。最后通过实验说明了该方法的有效性。
In order to reflect the decision ability 9f decision rule and improve the efficiency and classification accuracies in knowledge reduction, this paper developed a new method for incremental knowledge acquisition. This method introduced deci- sion information entropy to measure the uncertain degree of decision rules. The proposed method based on decision information entropy can be used to process dynamic data set without repeated computing, by updating the existed rule set step by step to improve the efficiency and classification accuracies in knowledge acquisition. Finally the experimental simulation shows that the proposed method is effective.
出处
《计算机应用研究》
CSCD
北大核心
2014年第4期989-992,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(61073146
61309014
61379114)
重庆市基础与前沿研究计划资助项目(cstc2013jcyj A40063)
中国与波兰政府间科技合作基金资助项目(国科外字[2010]179号)
关键词
决策熵
粗糙集
规则树
动态信息系统
增量式知识获取
规则提取
decision information entropy
rough set
rule tree
dynamic information systems
incremental knowledge acquisition
rule extraction