论文部分内容阅读
针对区间值信息系统基于粗糙集理论提出一种新的属性约简算法:首先计算同一属性下对象间的相似度,然后通过合取算子计算出所有属性下对象之间的相似度矩阵,再用模糊聚类中的传递闭包算子得到等价矩阵,将区间值信息系统转化为具有等价关系的信息系统并且进行约简,从而得到λ-核,同时给出了该算法的复杂度。最后通过一个实例表明这种算法的有效性和合理性。
A new attribute reduction algorithm is proposed based on rough set theory for interval valued information system. First, the similarity between objects under the same attribute is calculated. Then, the similarity matrix between objects under all attributes is calculated by combining operators. The equivalent matrices are obtained by the transfer closure operator in fuzzy clustering, and the information systems of interval value information are transformed into information systems with equivalent relation and reduced to get λ-kernel. The complexity of this algorithm is also given . Finally, an example shows the validity and rationality of this algorithm.