多属性单一趋势结构时序数据的聚类模型

Clustering model for multidimensional time series data of unitary trending structure

针对多属性单一趋势结构时序数据的特点,提出一种加权免疫遗传模糊C均值聚类方法。为确立相似度权值,建立权值优化模型,利用改进离子群算法对模型进行求解;针对传统模糊C均值初始中心敏感的问题,将免疫机理引入到遗传算法框架中,对模糊C均值进行改进。实例验证结果表明,权值优化模型是合理有效的,求解方法具有较高的收敛精度及速度,与其它方法相比,聚类方法具有较高的收敛精度。

According to the characteristic of multidimensional time series data of unitary trending structure,a weighted immune genetic fuzzy C-means clustering model was proposed.To clarify the similarity weights,weighted optimization model was established and the model was solved by using the improved particle swarm optimization algorithm.To overcome the problem that traditional fuzzy C-means algorithm is sensitive to initial center,the immune mechanism was introduced into the genetic framework to improve fuzzy C-means algorithm.The experimental results show that the weighted optimization model is reasonable and effective and the solving method has higher convergence precision and speed.The clustering method has higher convergence precision compared with other methods.