Orthogonal simple component analysis

A new, exploratory approach

Karim Anaya-Izquierdo, Frank Critchley, Karen Vines

Research output: Contribution to journalArticle

7 Citations (Scopus)
93 Downloads (Pure)

Abstract

Combining principles with pragmatism, a new approach and accompanying algorithm are presented to a longstanding problem in applied statistics: the interpretation of principal components. Following Rousson and Gasser "the ultimate goal is not to propose a method that leads automatically to a unique solution, but rather to develop tools for assisting the user in his or her choice of an interpretable solution." Accordingly, our approach is essentially exploratory. Calling a vector 'simple' if it has small integer elements, it poses the open question: What sets of simply interpretable orthogonal axes-if any-are angle-close to the principal components of interest? its answer being presented in summary form as an automated visual display of the solutions found, ordered in terms of overall measures of simplicity, accuracy and star quality, from which the user may choose. Here, 'star quality' refers to striking overall patterns in the sets of axes found, deserving to be especially drawn to the user's attention precisely because they have emerged from the data, rather than being imposed on it by (implicitly) adopting a model. Indeed, other things being equal, explicit models can be checked by seeing if their fits occur in our exploratory analysis, as we illustrate. Requiring orthogonality, attractive visualization and dimension reduction features of principal component analysis are retained.
Original languageEnglish
Pages (from-to)486-522
Number of pages37
JournalAnnals of Applied Statistics
Volume5
Issue number1
DOIs
Publication statusPublished - Mar 2011

Fingerprint

Stars
Principal Components
Star
Principal component analysis
Exploratory Analysis
Visualization
Display devices
Statistics
Dimension Reduction
Orthogonality
Unique Solution
Thing
Principal Component Analysis
Simplicity
Display
Choose
Angle
Integer
Model
Principal components

Keywords

  • stat.AP

Cite this

Orthogonal simple component analysis : A new, exploratory approach. / Anaya-Izquierdo, Karim; Critchley, Frank; Vines, Karen.

In: Annals of Applied Statistics, Vol. 5, No. 1, 03.2011, p. 486-522.

Research output: Contribution to journalArticle

Anaya-Izquierdo, Karim ; Critchley, Frank ; Vines, Karen. / Orthogonal simple component analysis : A new, exploratory approach. In: Annals of Applied Statistics. 2011 ; Vol. 5, No. 1. pp. 486-522.
@article{75f044d15bf24cbd9f050b440a37e941,
title = "Orthogonal simple component analysis: A new, exploratory approach",
abstract = "Combining principles with pragmatism, a new approach and accompanying algorithm are presented to a longstanding problem in applied statistics: the interpretation of principal components. Following Rousson and Gasser {"}the ultimate goal is not to propose a method that leads automatically to a unique solution, but rather to develop tools for assisting the user in his or her choice of an interpretable solution.{"} Accordingly, our approach is essentially exploratory. Calling a vector 'simple' if it has small integer elements, it poses the open question: What sets of simply interpretable orthogonal axes-if any-are angle-close to the principal components of interest? its answer being presented in summary form as an automated visual display of the solutions found, ordered in terms of overall measures of simplicity, accuracy and star quality, from which the user may choose. Here, 'star quality' refers to striking overall patterns in the sets of axes found, deserving to be especially drawn to the user's attention precisely because they have emerged from the data, rather than being imposed on it by (implicitly) adopting a model. Indeed, other things being equal, explicit models can be checked by seeing if their fits occur in our exploratory analysis, as we illustrate. Requiring orthogonality, attractive visualization and dimension reduction features of principal component analysis are retained.",
keywords = "stat.AP",
author = "Karim Anaya-Izquierdo and Frank Critchley and Karen Vines",
year = "2011",
month = "3",
doi = "10.1214/10-AOAS374",
language = "English",
volume = "5",
pages = "486--522",
journal = "Annals of Applied Statistics",
issn = "1932-6157",
publisher = "Institute of Mathematical Statistics",
number = "1",

}

TY - JOUR

T1 - Orthogonal simple component analysis

T2 - A new, exploratory approach

AU - Anaya-Izquierdo, Karim

AU - Critchley, Frank

AU - Vines, Karen

PY - 2011/3

Y1 - 2011/3

N2 - Combining principles with pragmatism, a new approach and accompanying algorithm are presented to a longstanding problem in applied statistics: the interpretation of principal components. Following Rousson and Gasser "the ultimate goal is not to propose a method that leads automatically to a unique solution, but rather to develop tools for assisting the user in his or her choice of an interpretable solution." Accordingly, our approach is essentially exploratory. Calling a vector 'simple' if it has small integer elements, it poses the open question: What sets of simply interpretable orthogonal axes-if any-are angle-close to the principal components of interest? its answer being presented in summary form as an automated visual display of the solutions found, ordered in terms of overall measures of simplicity, accuracy and star quality, from which the user may choose. Here, 'star quality' refers to striking overall patterns in the sets of axes found, deserving to be especially drawn to the user's attention precisely because they have emerged from the data, rather than being imposed on it by (implicitly) adopting a model. Indeed, other things being equal, explicit models can be checked by seeing if their fits occur in our exploratory analysis, as we illustrate. Requiring orthogonality, attractive visualization and dimension reduction features of principal component analysis are retained.

AB - Combining principles with pragmatism, a new approach and accompanying algorithm are presented to a longstanding problem in applied statistics: the interpretation of principal components. Following Rousson and Gasser "the ultimate goal is not to propose a method that leads automatically to a unique solution, but rather to develop tools for assisting the user in his or her choice of an interpretable solution." Accordingly, our approach is essentially exploratory. Calling a vector 'simple' if it has small integer elements, it poses the open question: What sets of simply interpretable orthogonal axes-if any-are angle-close to the principal components of interest? its answer being presented in summary form as an automated visual display of the solutions found, ordered in terms of overall measures of simplicity, accuracy and star quality, from which the user may choose. Here, 'star quality' refers to striking overall patterns in the sets of axes found, deserving to be especially drawn to the user's attention precisely because they have emerged from the data, rather than being imposed on it by (implicitly) adopting a model. Indeed, other things being equal, explicit models can be checked by seeing if their fits occur in our exploratory analysis, as we illustrate. Requiring orthogonality, attractive visualization and dimension reduction features of principal component analysis are retained.

KW - stat.AP

UR - http://www.scopus.com/inward/record.url?scp=84858306262&partnerID=8YFLogxK

UR - http://dx.doi.org/10.1214/10-AOAS374

UR - http://www.imstat.org

U2 - 10.1214/10-AOAS374

DO - 10.1214/10-AOAS374

M3 - Article

VL - 5

SP - 486

EP - 522

JO - Annals of Applied Statistics

JF - Annals of Applied Statistics

SN - 1932-6157

IS - 1

ER -