作者: Simona Korenjak-Černe , Vladimir Batagelj
DOI: 10.1007/978-3-642-56181-8_35
关键词:
摘要: The paper builds on the representation of units/clusters with a special type symbolic objects that consist distributions variables. Two compatible clustering methods are developed: leaders method, reduces large dataset to smaller set (clusters) which hierarchical method is applied reveal its internal structure. proposed approach illustrated USDA Nutrient Database.