Theory of Agglomerative Hierarchical Clustering

En savoir plus sur le livre

Recent theoretical advancements in agglomerative hierarchical clustering challenge the notion that its methodology is fully developed. The book presents a general algorithm that enhances traditional linkage methods, utilizing a recursive subprogram to create abstract tree structures for complex data sets. It proves a fundamental theorem for single linkage via fuzzy graphs and explores dendrogram reversals. Additionally, it introduces new methods using positive-definite kernels, making the theoretical insights applicable for users in practical scenarios.