本文提出了一种基于独立成分分析的改进k均值时间序列聚类算法,在独立成分分析对每个时间序列提取特征并使用动态时间规整算法对每个时间序列两两进行距离的计算之后,使用模拟退火算法结合k均值聚类算法寻找全局最优解,本文使用伪F统计量确定最优的聚类数目,最后,本文使用实际股票收益时间序列验证了该算法的可行性,得到了较好的结果。 In this paper, an improved k-means clustering algorithm based on independent component analysis is proposed. After the independent component analysis extracts the features of each time series and uses the dynamic time warping algorithm to calculate the distance of each time series, Simulated annealing algorithm combined with k-means clustering algorithm to find the global optimal solution, this paper uses the pseudo-F statistics to determine the optimal number of clusters. Finally, this paper verifies the feasibility of the algorithm using the actual stock returns time series, which is better the result of.