Time Series Segmentation and Clustering Method Based on Cloud Model

Abstract

In order to effectively evaluate the time series with high dimensionality and uncertainty, a time series segmentation and clustering method integrating cloud model analysis is proposed. The entropy and super entropy of the cloud model are used to identify the subsequence with the worst stability, further divide the subsequence, and dynamically realize segmentation to form a cloud model sequence. At the same time, the effectiveness evaluation indicator of segmented aggregation is constructed by the cloud model to determine the optimal number of segments. For different cloud model sequences, the cloud models are matched through the relationship of time window, and the similarity measures are given. The results of experiments indicate that the proposed method can effectively determine the number of time series segments, and greatly compress the original sequence, retain the basic characteristics of the data and improve the clustering efficiency of time series.