With the rapid development of science and technology, high-dimensional data have been widely used in various fields. Due to the complex characteristics of high-dimensional data, it is usually distributed in the union of several low-dimensional subspaces. In the past several decades, subspace clustering (SC) methods have been widely studied as they can restore the underlying subspace of high-dimensional data and perform fast clustering with the help of the data self-expressiveness property. The SC methods aim to construct an affinity matrix by the self-representation coefficient of high-dimensional data and then obtain the clustering results using the spectral clustering method. The key is how to design a self-expressiveness model that can reveal the real subspace structure of data. In this survey, we focus on the development of SC methods in the past two decades and present a new classification criterion to divide them into three categories based on the purpose of clustering, i.e., low-rank sparse SC, local structure preserving SC, and kernel SC. We further divide them into subcategories according to the strategy of constructing the representation coefficient. In addition, the applications of SC methods in face recognition, motion segmentation, handwritten digits recognition, and speech emotion recognition are introduced. Finally, we have discussed several interesting and meaningful future research directions.
With the development of science and technology, more and more data have been produced. For many of these datasets, only some of the data have labels. In order to make full use of the information in these data, it is necessary to classify them. In this paper, we propose a strong sparse quadratic kernel-free least squares semi-supervised support vector machine (), in which we add a norm regularization term to make it sparse. An NP-hard problem arises since the proposed model contains the norm and another nonconvex term. One important method for solving the nonconvex problem is the DC (difference of convex function) programming. Therefore, we first approximate the norm by a polyhedral DC function. Moreover, due to the existence of the nonsmooth terms, we use the sGS-ADMM to solve the subproblem. Finally, empirical numerical experiments show the efficiency of the proposed algorithm.
18 April 2023 in Statistical Analysis and Data Mining: An ASA Data Science Journal
The regularized clustering (RC) framework based on the fusion penalty has attracted extensive attention in the last decade because it does not require the prior knowledge of the number of clusters. Although the ground truth connections and weights among samples are beneficial for clustering, the performance of RC could be distorted by the omnipresent spurious connections in real-world data sets. To effectively address the issue, this letter constructs an enhanced regularized clustering model by incorporating a spurious connection detection mechanism into the objective function of the RC framework. The proposed model can effectively reduce the damage caused by spurious connections through adaptively identifying the importance of each connection. Furthermore, an alternating minimization algorithm is developed with detailed convergence analysis. Experimental results validate its effectiveness against several state-of-the-art RC methods
With the rapid growth of data volume and the increasing demand for real-time analysis, online subspace clustering has emerged as an effective tool for processing dynamic data streams. However, existing online subspace clustering methods often struggle to capture the complex and evolving distribution of such data due to their reliance on rigid dictionary learning mechanisms. In this paper, we propose a novel $\ell_0$ elastic net subspace clustering model by integrating the $\ell_0$ norm and the Frobenius norm, which owns the desirable block diagonal property. To address the challenges posed by the evolving data distributions in online data, we design a fast online alternating direction method of multipliers with an innovative dictionary update strategy based on support points, which are a set of data points to capture the underlying distribution of the data. By selectively updating dictionary atoms according to the support points, the proposed method can dynamically adapt to the evolving data characteristics, thereby enhancing both adaptability and computational efficiency. Moreover, we rigorously prove the convergence of the algorithm. Finally, extensive numerical experiments demonstrate that the proposed method improves clustering performance and computational efficiency, making it well-suited for real-time and large-scale data processing tasks.