一种基于Rough集获取规则知识的增量式学习方法

将Rough集理论应用于规则归纳系统,提出了一种基于粗糙集获取规则知识库的增量式学习方法,能够有效处理决策表中不一致情形,采用启发式算法获取决策表的最简规则,当新对象加入时在原有规则集基础上进行规则知识库的增量式更新,避免了为更新规则而重新运行规获取算法。并用UCI中多个数据集从规则集的规则数目、数据浓缩率、预测能力等指标对该算法进行了测试。实验表明了该算法的有效性。

Incremental Learning Method Based on Rough Set for Acquiring Rules Knowledge

A kind of incremental learning method based on rough sets is proposed here in rules induction system to acquire rule knowledge database.It deals with inconsistent examples,uses heuristic algorithm to gain concise rules which accord with decision table,and realizes incremental methodology for updating the existing decision rules when a new example is added in decision table.At last,this paper uses some standard datasets of UCI database to test this algorithm in some aspects such as rules count of the rules set,data enriching,forecast accuracy of the rules set.Experimentation makes sure validity of this new algorithm.