A validation methodology in hierarchical clustering

17 May, 2005·
Jorge N. Tendeiro
Jorge N. Tendeiro
,
Fernanda Sousa
Abstract
This paper presents a validation methodology in ascending hierarchical clustering. The objects in validation are clustering hierarchies, and simulation is used. Under certain conditions, this methodology allows us to evaluate the quality of hierarchical structures, its robustness and fiability, according to the data structure. The effect of the application of a given criterion on some kind of structures is also analyzed.
Location

Brest

Brest

talks
Jorge N. Tendeiro
Authors
Full Professor
Jorge N. Tendeiro is a professor at the Graduate School of Advanced Science and Engineering, School of Informatics and Data Science, Hiroshima University, Japan. His main research interest is in the Bayesian statistical framework. He is also interested in item response theory in general and person fit analysis in particular.