Abstract: Attribute reduction that has been proved to be a NP-hard problem is one of the important issues of the KDD based on the rough set theory. A novel attribute reduction algorithm of rough set based on the vertebrate immune mechanism is proposed. The main operators of the algorithm include memory cells producing, clone selection, hyper-mutation and population updating. The key of its design is to integrate discernible ability and the elements in the condition attribute set into one unified affinity maturation objection. The different attribute reduction sets that can maintain the ability of classification can be found through maintaining the diversity of antibody population with renewal of antibody and similar antibodies suppression. The experimental results show that it is effective.

Key words: immune mechanism, rough set, attribute reduction

摘要： 在基于粗糙集理论的知识发现中，属性约简是其中重要的研究内容之一，已经被证明是NP完全问题。基于生物免疫原理，提出了一种新型粗糙集属性约简算法。该算法由记忆细胞获取、克隆选择、超变异和群体更新4种算子构成。算法设计的重点在于将分类精度和约简中所含属性个数集成为一个统一的亲合度成熟目标，并通过抗体更新和抗体相似性抑制来维持群体的多样性，以获得多个符合分类质量要求的属性约简集。实验结果证明了该算法的有效性。

关键词: 免疫原理, 粗糙集, 属性约简

CLC Number: 

• TP18