In real life there are many kinds of phenomena that are better described by interval bounds than by single-valued variables. In fact, intervals take into account the imprecision due to measurement errors. When there is information about the imprecision distribution the fuzzy data coding is used to represent the imprecision. In this paper, we first review the main dimension reduction techniques for interval-valued data and then we propose a midpoints and radii-based approach. In particular, an alternative pre-processing and Procrustean rotation of the traditional midpoints and radii approach is proposed.

Principal component analysis for interval data: common approaches and variations / IODICE D'ENZA, Alfonso; Schisa, Vivana; Palumbo, Francesco. - In: STATISTICA APPLICATA. - ISSN 2038-5587. - 33:3(2021), pp. 249-270. [10.26398/IJAS.0033-013]

Principal component analysis for interval data: common approaches and variations

Iodice D'Enza Alfonso
;
Palumbo Francesco
2021

Abstract

In real life there are many kinds of phenomena that are better described by interval bounds than by single-valued variables. In fact, intervals take into account the imprecision due to measurement errors. When there is information about the imprecision distribution the fuzzy data coding is used to represent the imprecision. In this paper, we first review the main dimension reduction techniques for interval-valued data and then we propose a midpoints and radii-based approach. In particular, an alternative pre-processing and Procrustean rotation of the traditional midpoints and radii approach is proposed.
2021
Principal component analysis for interval data: common approaches and variations / IODICE D'ENZA, Alfonso; Schisa, Vivana; Palumbo, Francesco. - In: STATISTICA APPLICATA. - ISSN 2038-5587. - 33:3(2021), pp. 249-270. [10.26398/IJAS.0033-013]
File in questo prodotto:
File Dimensione Formato  
11_IJAS_PCA_int.pdf

accesso aperto

Tipologia: Versione Editoriale (PDF)
Licenza: Non specificato
Dimensione 1.46 MB
Formato Adobe PDF
1.46 MB Adobe PDF Visualizza/Apri

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/11588/883430
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact