The topic of this tutorial is the effective dimensionality (ED) of a dataset, that is, the equivalent number of orthogonal dimensions that would produce the same overall pattern of covariation. The ED quantifies the total dimensionality of a set of variables, with no assumptions about their underlying structure. The ED of a dataset has important implications for the “curse of dimensionality”; it can be used to inform decisions about data analysis and answer meaningful empirical questions. The tutorial offers an accessible introduction to ED, distinguishes it from the related but distinct concept of intrinsic dimensionality, critically reviews various ED estimators, and gives indications for practical use with examples from personality research. An R function is provided to implement the techniques described in the tutorial.

Effective Dimensionality: A Tutorial

Del Giudice, Marco
2021-01-01

Abstract

The topic of this tutorial is the effective dimensionality (ED) of a dataset, that is, the equivalent number of orthogonal dimensions that would produce the same overall pattern of covariation. The ED quantifies the total dimensionality of a set of variables, with no assumptions about their underlying structure. The ED of a dataset has important implications for the “curse of dimensionality”; it can be used to inform decisions about data analysis and answer meaningful empirical questions. The tutorial offers an accessible introduction to ED, distinguishes it from the related but distinct concept of intrinsic dimensionality, critically reviews various ED estimators, and gives indications for practical use with examples from personality research. An R function is provided to implement the techniques described in the tutorial.
2021
29-mar-2020
Pubblicato
https://www.tandfonline.com/doi/full/10.1080/00273171.2020.1743631
File in questo prodotto:
File Dimensione Formato  
DelGiudice_2020_effective-dimensionality_tutorial_pre.pdf

Open Access dal 30/03/2021

Tipologia: Bozza finale post-referaggio (post-print)
Licenza: Creative commons
Dimensione 2.23 MB
Formato Adobe PDF
2.23 MB Adobe PDF Visualizza/Apri
3145866-2-18.pdf

Accesso chiuso

Tipologia: Documento in Versione Editoriale
Licenza: Copyright Editore
Dimensione 9.06 MB
Formato Adobe PDF
9.06 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11368/3066297
Citazioni
  • ???jsp.display-item.citation.pmc??? 9
  • Scopus 15
  • ???jsp.display-item.citation.isi??? 18
social impact