Multi-view reduced dimensionality K-means clustering with σ-norm and Schatten p-norm

Xiangdong Zhang, Fangfang Li, Zhaoyang Shi, Ming Yang

Abstract: Highlights•In order to avoid the influence of dimensional curse and redundant features in the original space, we use dimension reduction technology to process high-dimensional multi-view data.•We use the σ<math><mi is="true">σ</mi></math>-norm as an adaptive loss minimization, it flexibly weighs sparsity and continuity.•We form the label matrix into a third-order tensor and apply the Schatten p-norm to mine higher-order relationships between multiple views.•An optimization strategy based on ALM is proposed. A large number of experiments show that our method has good performance.