Abstract:Localized incomplete multiple kernel k -means (LI-MKKM) is recently put forward to boost the clustering accuracy via optimally utilizing a quantity of prespecified incomplete base kernel matrices. Despite achieving significant achievement in a variety of applications, we find out that LI-MKKM does not sufficiently consider the diversity and the complementary of the base kernels. This could make the imputation of incomplete kernels less effective, and vice versa degrades on the subsequent clustering. To tackle these problems, an improved LI-MKKM, called LI-MKKM with matrix-induced regularization (LI-MKKM-MR), is proposed by incorporating a matrix-induced regularization term to handle the correlation among base kernels. The incorporated regularization term is beneficial to decrease the probability of simultaneously selecting two similar kernels and increase the probability of selecting two kernels with moderate differences. After that, we establish a three-step iterative algorithm to solve the corresponding optimization objective and analyze its convergence. Moreover, we theoretically show that the local kernel alignment is a special case of its global one with normalizing each base kernel matrices. Based on the above observation, the generalization error bound of the proposed algorithm is derived to theoretically justify its effectiveness. Finally, extensive experiments on several public datasets have been conducted to evaluate the clustering performance of the LI-MKKM-MR. As indicated, the experimental results have demonstrated that our algorithm consistently outperforms the state-of-the-art ones, verifying the superior performance of the proposed algorithm.

Localized Multiple Kernel Learning With Dynamical Clustering and Matrix Regularization.

Generalized Convexity-Based Inexact Projection Method for Multiple Kernel Learning

A Unified View of Localized Kernel Learning

Localized Multiple Kernel Learning Via Sample-Wise Alternating Optimization

Probability-confidence-kernel-based localized multiple kernel learning with l p norm

Multiple Kernel Clustering via Local Regression Integration

Localized Incomplete Multiple Kernel k-Means With Matrix-Induced Regularization

Multiple Kernel Subspace Clustering with Local Structural Graph and Low-Rank Consensus Kernel Learning

Consensus Local Graph for Multiple Kernel Clustering

Multiple Kernel K-Means Clustering with Matrix-Induced Regularization

Local Sample-Weighted Multiple Kernel Clustering With Consensus Discriminative Graph

Local kernels based graph learning for multiple kernel clustering

Multiple Empirical Kernel Learning with Locality Preserving Constraint

On the Impact of Regularization Variation on Localized Multiple Kernel Learning.

Localized Multiple Kernel Learning for Realistic Human Action Recognition in Videos

Multiple empirical kernel learning based on local information

Multiple Kernel Subspace Learning for Clustering and Classification

Simultaneous Global and Local Graph Structure Preserving for Multiple Kernel Clustering

Learning Local Graph from Multiple Kernels

Joint diversity regularization and graph regularization for multiple kernel k-means clustering via latent variables.

A Unified Framework for Discrete Multi-kernel K-Means with Kernel Diversity Regularization