基于人工免疫和基因表达式编程的多维复杂关联规则挖掘方法

为满足复杂数据挖掘应用对处理丰富语义的要求，引入了多维复杂关联规则概念，提出了通过人工免疫循环控制的基因表达式编程挖掘方法。构造了有特色的抗体和免疫细胞结构，能有效减少计算量；设计了特有的否定选择策略，可以消除无用的和冗余的免疫细胞；引出了逆否规则与原规则同为强规则的启发式过滤准则，可有效约简规则数目。实验表明，新方法能够高效、准确地挖掘多维复杂关联规则；在一定条件下，新方法的否定选择策略可将挖掘效率提高达1~3个数量级。

Mining Multi-dimensional Complex Association Rule Based onArtificial Immune System and Gene Expression Programming

In order to handle rich semantics for complex data mining application,the formal concept of Multi-dimensional Complex Association Rule(MDCAR) was proposed.To mine it,a novel method based on Artificial Immune Gene Expression Programming (AIGEP) was introduced,where,new structures of antibody and immune cell were designed to decrease computing complexity,the special negative select strategy was presented to eliminate invalid or redundant immune cells according to system requirements,and a heuristic MDCAR reduction criterion was introduced,that is,a strong rule is fine only if the contra-positive of it is strong.Experiments showed that the new method can mine MDCAR with good efficiency and high precision and improve the performance,in certain case,10~1000 times higher than that without negative select strategy.