基于多视图的多元时间序列聚类方法
收藏资源简介:
基于多视图的多元时间序列聚类方法(Multi-view Multi-variate Time Series Clustering Method,CSMVC)面向特征维度高、结构复杂等特点从数据表示的角度出发,挖掘数据的多视图信息。CSMVC利用了多元时间序列数据的多视图特性,采用多视图聚类方法来提高聚类能力。采用一种新型的NMF方法,实现了从不同角度提取数据的全局结构和局部结构。
The Multi-view Multi-variate Time Series Clustering Method (CSMVC) targets datasets with high feature dimensionality and complex structures. Starting from the perspective of data representation, it excavates the multi-view information inherent in the data. CSMVC leverages the multi-view properties of multivariate time series data, and employs multi-view clustering techniques to improve clustering performance. Specifically, it utilizes a novel non-negative matrix factorization (NMF) method to extract both the global and local structures of the data from diverse perspectives.




