Incremental Reducts Based on Nearest Neighbor Relations and Linear Classifications
Incremental Reducts Based on Nearest Neighbor Relations and Linear Classifications
复制标题
基于最近邻关系和线性分类的增量归约
DOI：
10.1109/iiai-aai.2019.00113
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
T. Matsuo
中科院分区：
文献类型：
--
作者：
N. Ishii;Ippei Torii;K. Iwata;Kazuya Odagiri;Toyoshiro Nakashima;T. Matsuo
数据的降维或变量约简是一个重要的问题，在应用领域中对高维数据的分析需要进行降维。粗糙集理论是将高维数据约简为低维数据进行分类的基础。提出了一种基于最近邻关系和线性分类的增量约简生成方法。首先，最近邻关系的近似约简中起着重要的作用。其次，在退化凸锥上生成完全约简，并在其中进行边操作。最后，利用线性分类和凸锥上的最近邻关系生成增量约简。
Dimension or variables reduction of data is an important problem and the reduction is needed for the analysis of higher dimensional data in the application domain. Rough set is fundamental and useful to reduce higher dimensional data to lower one for the classification. We develop a generation method of incremental reducts based on nearest neighbor relations and linear classifications using added data. First, the nearest neighbor relation is shown to play a fundamental role for the approximated reducts. Next, the complete reducts are generated on the degenerate convex cones, in which edge operations are performed. Finally, the incremental reducts are generated using the linear classification and the nearest neighbor relations on the convex cones.