摘要
提出基于非线性Kdv-KSV方程平衡泛函的优化聚类算法,运用非线性kdv-ksv方程式定义映射集范数设置初始类聚中心,通过泛函空间完成聚类数据的中心向量数域计算,划分聚类数据目标函数,利用隶属矩阵判断划分得到最优聚类的过程。仿真实验表明,基于非线性Kdv-KSV方程平衡泛函的优化聚类算法,数据收敛速度更快,动态特性跟踪效果更好,并且降低了聚类计算对初始值的依赖性。提高了处理高维数据的能力。
Put forward an algorithm based on nonlinear Kdv- KSV balance functional optimization clustering, the use of nonlinear Kdv KSV equations define the mapping set norm set initial type center, through the functional space data cluster-ing number field calculation, the center of the vector data partitioning clustering objective function, using the membership matrix judgment to get the optimal clustering process. Simulation results show that, based on the nonlinear equation of Kdv-KSV balance functional optimization clustering algorithm, data faster convergence speed, dynamic tracking effect is better, and reduces the cluster computing dependence on initial value. Improve the ability of dealing with high-dimension-al data.
出处
《科技通报》
北大核心
2015年第12期121-122,138,共3页
Bulletin of Science and Technology
