Approaches to automated grouping in singular spectrum analysis are considered. A new method for the identification of periodic components is proposed. The possibilities of extensions to multivariate time series and images are discussed.

翻译：本文讨论了奇异谱分析中自动分组的方法。提出了一种用于识别周期分量的新方法。还讨论了将方法推广到多元时间序列和图像的可能性。