属性值变化时变精度粗糙集模型中近似集动态更新的矩阵方法研究

摘要

In order to study the effective methods of dynamic knowledge update of information system from the view of matrix,this paper proposed a matrix-based incremental updating method for the lower and upper approximations of a concept in variable precision rough set while the attribute values evolves over time.Then it designed the corresponding algorithms.Furthermore,it developed the programs for the proposed matrix-based incremental algorithm of a concept and the matrix-based non-incremental algorithms on MATLAB.Finally,the performance test of two kinds of algorithms carried out on UCI datasets.The result of the performance test demonstrates the feasibility,conciseness and validity of the proposed matrix-based incremental method for updating approximations.%为了从矩阵的视角探讨信息系统动态知识更新的有效方法和途径,提出了信息系统的属性值发生变化时变精度粗糙集模型中概念上、下近似集增量式更新的矩阵方法并构造出相应的算法；随后在MATLAB平台上开发出近似集增量式更新和非增量式更新的两类矩阵算法的程序；最后在UCI数据集上进行两种矩阵算法的性能测试.测试结果表明,概念近似集增量式更新矩阵算法的可行、简洁和高效.