Distance Stability Analysis in Multidimensional Scaling using Jackknife

J. F. Vera Vera

Stability or sensibility analysis is an important topic in data analysis that has received little attention in the application of Multidimensional Scaling, for which the only available approaches are given in terms of a coordinate-based analytical jackknife methodology. Although in MDS the prime interest is in rating the stability of the points in the configuration, this methodology may be influenced by imprecisions derived from the inherent required Procrustean method. In this work, an analytical distance-based jackknife procedure is proposed to study stability and cross-validation in MDS in terms of the jackknife distances, which are not influenced by any Procrustes procedure. Stability and cross-validation are analyzed in terms of the MDS distances instead of the distances between procrustean configurations, and a stability-relevant configuration is given related to the best approximation of the averaged jackknife-derived MDS distances in terms of Euclidean distances.

Palabras clave: Multidimensional Scaling, Jackknife, Stability, Analysis of Dispersion, Cross-Validation

M05.2 Grupo de Análisis Multivariante y Clasificación III
6 de septiembre de 2016  11:00
0.09 - Aula de proyectos 2

A. Satorra

J. Fortiana Gregori, A. Esteve Gómez

J. L. Vicente Villardón