论文部分内容阅读
针对变精度粗糙集要依据先验领域知识来确定变精度阈值,大大影响属性约简和分类效率等问题,本文将可拓学和粗糙集应用到熵理论中,并结合不完备知识系统属性约简特征,提出一种变精度阈值的可拓熵属性约简算法(RAEE),该算法充分利用可拓熵在处理动态变化数据方面优势,能动态确定因先验知识不足的变精度阈值β,并选取最优阈值区间确定属性间近似依赖为启发式标准来进行可拓熵β-近似属性约简.仿真实验结果表明,RAEE算法在不完备系统中属性约简效率较高且对噪音具有较好抗干扰性.
For the variable precision rough set to determine the variable precision threshold based on the prior knowledge of the field, greatly affecting the attribute reduction and classification efficiency problems, this paper applies extension theory and rough set to the entropy theory, combined with the properties of the incomplete knowledge about (RAEE), this algorithm makes full use of the advantages of entropy in dealing with dynamic change data, and can dynamically determine the variable precision threshold β due to the prior knowledge, And select the optimal threshold interval to determine the approximate dependency between attributes to determine the heuristic entropy reduction entropy attribute reduction.Experimental results show that the RAEE algorithm in the incomplete system attribute reduction efficiency is high and the noise is more Good anti-interference.