基于最邻近聚类和向量模糊c-均值的混沌预测

Prediction of chaos based on the nearest neighbor clustering and vector fuzzy c-means clustering

针对混沌时间序列难预测的问题,提出一种新的基于最邻近聚类和向量模糊c-均值(FCMV)聚类算法的模糊建模方法。其前提参数辨识分两步,首先用最近邻聚类法初始划分输入空间,得到规则数及初始聚类中心,再用FCMV把具有相同收敛向量的聚类中心归到同一个区域来优化前一步得到的聚类中心,得到前提参数;采用递推最小二乘算法辨识模型的结论参数。最后通过对Mackey-Glass混沌时间序列的建模和预测验证了该方法的有效性与实用性。

A new method for fuzzy modeling based on a nearest neighbor clustering and vector fuzzy c-means algorithm （FCMV） is presented. The premise parameter identification consists of two steps： first, an initial fuzzy partition of input space by a nearest neighbor clustering method is performed to get the number of rules and the initial clustering center, then the initial clustering centers with the same convergent vector are grouped into the same region using the FCMV algorithm, thus the premise parameters are got. The conclusion parameters are identified by the recursive least square. At last the proposed method is applied to the modeling and prediction of the chaotic Mackey-Glass time series, and the results demonstrate the effectiveness and practicability of the method.