Minimum Description Length based Granular-Ball Tree Regularization for Spectral Clustering
Researchers have developed a new spectral clustering method called MDL-GBTRSC, which aims to improve the construction of affinity graphs. This method utilizes a Minimum Description Length (MDL) principle to build a granular-ball tree, effectively regularizing the sample-level graph. By preserving reliable local connectivity and using stable leaf balls for coding-scale information, MDL-GBTRSC connects representation learning with graph construction. Experiments indicate that this approach outperforms existing spectral clustering methods on various datasets. AI
IMPACT Introduces a novel approach to spectral clustering, potentially improving data analysis and representation learning in machine learning applications.