Componentes principales esféricas y matriz de covariancia de determinante mínimo: una aplicación sobre indicadores de carencias críticas
Principal Components Analysis (PCA) is widely used in multivariate statistical analysis. The objective of this method is to represent a set of n observations with p variables through a smaller number of variables that are linear combinations of the original ones, keeping as much as possible the orig...
Guardado en:
| Autores principales: | , |
|---|---|
| Otros Autores: | |
| Formato: | conferenceObject documento de conferencia acceptedVersion |
| Lenguaje: | Español |
| Publicado: |
2017
|
| Materias: | |
| Acceso en línea: | http://hdl.handle.net/2133/7611 http://hdl.handle.net/2133/7611 |
| Aporte de: |
| Sumario: | Principal Components Analysis (PCA) is widely used in multivariate statistical analysis. The objective of this method is to represent a set of n observations with p variables through a smaller number of variables that are linear combinations of the original ones, keeping as much as possible the original variability in the data. Two robust methods are presented in this work: the Minimum Covariance Determinant method (MCD) and the Spherical Principal Components method (SPC). The objective of this work is to compare these two methods with the classic PCA when applied to data related to indicators of critical needs in cities with more than 2000 inhabitants in the province of Santa Fe. The data comes from the National Census from 2010. In order to summarize the differences among the cities it is necessary to consider a greater number of principal components in the robust methods than in the classic method. The reason is that the latter uses variability measures that are influenced by outliers while the robust methods use a solid measure that is free from this problem |
|---|