基于鲁棒模糊聚类的混沌时间序列预测

采用一种基于鲁棒模糊聚类算法的模糊辨识方法，通过引入局部划分关联度因子，增强了系统辨识的抗干扰能力，提高了系统辨识的鲁棒性.首先用最近邻模糊聚类法划分初始输入空间，得到模糊规则数及初始聚类中心；然后用鲁棒模糊聚类算法求解并优化模糊隶属度和聚类中心，建立高精度的T-S模糊模型；最后利用最小二乘法辨识模型的初始结论参数，进一步利用带遗忘因子的递推最小二乘法优化结论参数.采用该方法对Mackey-Glass混沌时间序列进行建模和预测，仿真结果表明利用本方法可以进行准确建模和预测，验证了本方法的鲁棒性、有效性和实 关键词： 最近邻模糊聚类 鲁棒模糊聚类 混沌时间序列 最小二乘法

Prediction of chaotic time series based on robust fuzzy clustering

We propose a new method for fuzzy modeling based on a robust fuzzy-clustering algorithm. The induced local spatial similarity improved the system's robustness to noise and outsider and predicted the robustness of the modeling system. Starting from an initial fuzzy partition of input space by a nearest-neighbor clustering method to get the number of rules and the initial clustering center, we can compute and optimize the fuzzy membership and the clustering center with a robust fuzzy-clustering algorithm and get the high precision T-S model. The obtained parameters were identified by the least square method and further optimized by selective recursive least square. The proposed method was applied to simulations of chaotic Mackey-Glass time series modeling and prediction. The results demonstrated the robustness, effectiveness and practicability of the method.