一种变粒度的规则提取算法

发布时间：2018-05-13 20:19

  本文选题：粗糙集 + 属性约简　； 参考：《重庆邮电大学学报(自然科学版)》2016年06期

【摘要】：属性约简和值约简是粗糙集理论中知识获取的重要组成部分。通常,在知识获取的过程中先进行属性约简,然后在其基础上进行规则提取。但在实际应用中,属性约简在简化信息系统与提高规则提取效率的同时,原始信息系统中有些重要的条件属性可能被丢弃,从而导致属性约简后对信息系统进行知识获取得到的规则其数量与简化程度并不占优。针对上述问题,提出一种基于粒度变化的规则获取算法,通过属性粒度从粗到细的变化,直接从原始信息系统中提取规则;采用该方法得到的规则与属性约简后得到的规则相比,它们的数量与平均每条规则包含的特征属性数相对较少。最后,在理论分析的基础上,通过实例验证了算法可行性,并通过实验验证了算法的正确性和高效性。
[Abstract]:Attribute reduction and value reduction are important parts of knowledge acquisition in rough set theory. Usually, attribute reduction is performed in the process of knowledge acquisition, and then rule extraction is carried out on the basis of attribute reduction. However, in practical application, attribute reduction can simplify information system and improve the efficiency of rule extraction, while some important conditional attributes in the original information system may be discarded. As a result, the number and simplification of the rules obtained from the knowledge acquisition of the information system after attribute reduction are not dominant. Aiming at the above problems, a rule acquisition algorithm based on granularity change is proposed, which extracts the rules directly from the original information system by changing the attribute granularity from coarse to fine. Compared with the rules obtained by attribute reduction, the number of rules obtained by this method is less than the average number of characteristic attributes contained in each rule. Finally, on the basis of theoretical analysis, the feasibility of the algorithm is verified by an example, and the correctness and efficiency of the algorithm are verified by experiments.
【作者单位】： 重庆邮电大学计算智能重庆市重点实验室;重庆邮电大学理学院;
【基金】：国家自然科学基金项目(61472056;61309014) 重庆邮电大学科研训练计划项目(A2014-45)~~
【分类号】：TP181
，

本文编号：1884663