摘要
为了提高蚂蚁矿工算法的性能,使蚂蚁矿工算法能够更好处理具有连续离散属性的数据集,采用混合正态核函数处理连续属性。使用混合域蚂蚁矿工算法来发现无序的规则集,这些规则解释了训练集中隐藏的知识,同时能够将给定的数据集进行分类。实验结果表明,混合域蚂蚁矿工算法能产生可以接受的预测准确率,发现更简单的规则,在故障诊断系统中采用混合域蚂蚁矿工算法产生的规则比蚂蚁矿工算法更有效。
To improve Ant-miner so that it can process with continuous attributes directly,mixture of normal kernel probability density function(PDF) is applied to discover rule sets for mixed variables in a data set.Ant-miner for mixed variables is proposed to discover sets of unstructured classification rules.These rules are interpreted as knowledge hidden in the training data and are also used to classify the given data.Experiments show that our method has an acceptable predictive accuracy while discovering rules that are simpler and more comprehensive.In fault diagnosing systems,ant-miner for mixed variables is more efficient than the results of ant-miner.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第3期1055-1058,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(50875131)
南京农业大学青年科学基金项目(0601)
关键词
数据挖掘
分类规则
蚁群优化
连续域优化
概率密度函数
data mining
classification rules
ant colony optimization
continuous optimization
probability density function