Multiple Kernel Clustering

Multiple Kernel Clustering (MKC) aims to improve clustering accuracy by combining information from multiple kernel functions, each capturing different aspects of the data's non-linear structure. Current research focuses on developing efficient algorithms, such as those based on factorization and graph-based methods, to handle large datasets and optimize kernel weighting strategies, including integrating both correlation and dissimilarity measures. These advancements enhance the applicability of MKC across diverse fields, from improving model selection in domain adaptation to analyzing complex time-series data in psychopathology and enabling more interpretable clustering results through decision tree approximations.