论文部分内容阅读
Lingras提出的粗K均值聚类算法易受随机初始聚类中心和离群点的影响,可能出现一致性和无法收敛的聚类结果.对此,提出一种改进的粗K均值算法,选择潜能最大的K个对象作为初始的聚类中心,根据数据对象与聚类中心的相对距离来确定其上下近似归属,使边界区域的划分更合理.定义了广义分类正确率,该指标同时考虑了下近似集和边界区域中的对象,评价算法性能更准确.仿真实验结果表明,该算法分类正确率高,收敛速度快,能够克服离群点的不利影响.
The rough K-means clustering algorithm proposed by Lingras is susceptible to random initial cluster centers and outliers, which may result in consistent and non-convergent clustering results. In this regard, an improved rough K-means algorithm is proposed to select the potential According to the relative distance between the data object and the cluster center, the maximum K objects are determined as the upper and lower approximate attributions, so that the division of the boundary area is more reasonable. The generalized classification accuracy rate is defined, which considers both Approximation sets and objects in the boundary region, the performance of the proposed algorithm is more accurate.The simulation results show that the proposed algorithm has high classification accuracy and fast convergence, and can overcome the adverse effects of outliers.