基于压缩理论的区间概念格参数优化模型

发布时间：2018-08-22 19:02

【摘要】：在由形式背景构建区间概念格之前,首先要确定区间参数[α,β],区间参数的选取影响着概念外延、格结构以及提取的关联规则数量和精度。为了获取区间概念格的压缩度达到最大时的[α,β],首先,提出了基于形式背景的二元关系对的相似度和二元关系上的覆盖近邻空间的定义,得到二元关系对的相似矩阵,并根据γ相似类求得的覆盖来计算二元关系对的近邻;其次,给出基于参数变化的概念集合更新算法,在非重建的基础上得到各区间参数下概念集合,并结合各区间参数下二元关系对的近邻空间,进一步构建基于压缩理论的区间概念格参数优化模型,依据压缩度的大小以及变化趋势寻找区间参数最优值;最后,通过实例验证了模型的有效性。
[Abstract]:Before constructing interval concept lattice from formal background, the interval parameters [伪, 尾] should be determined first. The selection of interval parameters affects the concept extension, lattice structure, the number and accuracy of extracted association rules. In order to obtain [伪, 尾] when the contraction degree of interval concept lattice reaches the maximum, the similarity degree of binary relation pair based on formal background and the definition of covering nearest neighbor space on binary relation are proposed, and the similarity matrix of binary relation pair is obtained. The nearest neighbor of binary relation pair is calculated according to the cover obtained by 纬 similar class. Secondly, the concept set updating algorithm based on parameter variation is given, and the concept set under each interval parameter is obtained on the basis of non-reconstruction. Combined with the adjacent space of binary relation pair under each interval parameter, the optimization model of interval concept lattice parameter based on compression theory is further constructed, and the optimal value of interval parameter is found according to the magnitude of compression degree and the change trend. Finally, The validity of the model is verified by an example.
【作者单位】： 华北理工大学理学院;河北省数据科学与应用重点实验室;
【基金】：国家自然科学基金资助项目(61370168,61472340) 河北省自然科学基金资助项目(F2016209344) 华北理工大学青年科学研究基金资助项目(Z201517)~~
【分类号】：TP301.6
，

本文编号：2198024