Adaptation of interval PCA to symbolic histogram variables

Authors:
Sun Makosso-Kallyth;Edwin Diday
Affiliations:
CRCHUQ, Plate forme recherche clinique, Quebec, Canada;Université Paris Dauphine, CEREMADE, Paris Cedex 16, France 75775
Venue:
Advances in Data Analysis and Classification
Year:
2012

Citing 5
Cited 0

The statistical analysis of compositional data

The statistical analysis of compositional data
Analysis of Symbolic Data: Exploratory Methods for Extracting Statistical Information from Complex Data

Analysis of Symbolic Data: Exploratory Methods for Extracting Statistical Information from Complex Data
Symbolic Data Analysis: Conceptual Statistics and Data Mining (Wiley Series in Computational Statistics)

Symbolic Data Analysis: Conceptual Statistics and Data Mining (Wiley Series in Computational Statistics)
Symbolic Data Analysis and the SODAS Software

Symbolic Data Analysis and the SODAS Software
The quantile method for symbolic principal component analysis

Statistical Analysis and Data Mining

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper is an adaptation of symbolic interval Principal Component Analysis (PCA) to histogram data. We proposed two methodologies. The first one involved three steps: the coding of bins of histogram, the ordinary PCA of means of variables and the representation of dispersion of symbolic observations we call concepts. For the representation of dispersion of these concepts we proposed the transformation of histograms into intervals. Then, we suggest the projection of the hypercubes or the interval lengths associated to each concept on the principal axes of the ordinary PCA of means. In the second methodology, we proposed the use of the three previous steps with the angular transformation.