摘要
针对权值约束、相互制约的两类数据A、B分类问题,提出一种带修补规则的数据分割算法。首先,基于密度法,依赖于数据类B确定数据类A的代表点,并依据最短距离原则和数据类A的代表点,对数据类B进行分类;进而,依据邻域法的思想及数据类B的权值,对各个分类进行修补,使各类满足权值约束;最后,基于优化模型,对数据类B的分类进行优化。实验结果表明,该分割算法获得了满意的分类效果。
A repair rule-based data-segment algorithm is proposed for the classification problem of two kinds of data (A and B) with mutual restriction and the weight constraints, First, the data B is divided into several sub-classes related to the density method, which is based on the representative points of the data A and the shortest distance rule. Second, these sub-classes are repaired to satisfy the weight constraints depending upon the idea of the neighborhood method and the weights of the data B; on the other hand, they are optimized based on the proposed optimization model. Numerical experiments show that the algorithm can achieve the satisfactory classification effect.
出处
《贵州大学学报(自然科学版)》
2009年第2期85-89,共5页
Journal of Guizhou University:Natural Sciences
基金
教育部科学技术研究重点项目(208125)
关键词
聚类
分类
权值约束
密度
优化
clustering, classification, weight constraint, density, optimization.