Kernel Spectral Clustering

Kernel spectral clustering is a machine learning technique that uses kernel methods to perform spectral clustering on data, effectively handling non-linear relationships by mapping data into a higher-dimensional feature space. Current research focuses on improving the robustness and efficiency of these methods, particularly for large datasets and data with noise or outliers, through techniques like sparse kernel representations and optimized algorithms that avoid computationally expensive matrix operations. These advancements are significant because they enable the application of kernel spectral clustering to larger and more complex datasets, improving performance in various applications such as image segmentation and node clustering in complex networks.

Papers

May 27, 2024

HeNCler: Node Clustering in Heterophilous Graphs through Learned Asymmetric Similarity
Sonny Achten, Francesco Tonin, Volkan Cevher, Johan A. K. Suykens
Heterophilic Graph Node Clustering Asymmetric Similarity Kernel Spectral Clustering

January 30, 2024

Robust Kernel Sparse Subspace Clustering
Ivica Kopriva
Subspace Clustering Sparse Kernel Kernel Spectral Clustering

October 20, 2023

Accelerated sparse Kernel Spectral Clustering for large scale data clustering problems
Mihaly Novak, Rocco Langone, Carlos Alzate, Johan Suykens
Low Rank Approximation Kernel Matrix Kernel Principal Component Analysis Kernel Spectral Clustering

Kernel Spectral Clustering

Papers

HeNCler: Node Clustering in Heterophilous Graphs through Learned Asymmetric Similarity

Robust Kernel Sparse Subspace Clustering

Accelerated sparse Kernel Spectral Clustering for large scale data clustering problems