Alternating least squares in nonlinear principal components

Masahiro Kuroda, Yuichi Mori, Masaya Iizuka, Michio Sakakihara

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Principal components analysis (PCA) is probably the most popular descriptive multivariate method for analyzing quantitative data with ratio and interval scale measures. When applying PCA to nominal and ordinal data, the data are processed by a method such as optimal scaling, which nonlinearly transforms nominal and ordinal data into quantitative data. Therefore, PCA with optimal scaling is called nonlinear PCA. Nonlinear PCA reveals nonlinear relationships among variables with different measurement levels and therefore presents a more flexible alternative to ordinary PCA. The alternating least squares algorithm is utilized for nonlinear PCA. The algorithm alternates between optimal scaling for quantifying nominal and ordinal data and ordinary PCA for analyzing optimally scaled data. This article discusses two nonlinear PCA algorithms, namely, PRINCIPALS and PRINCALS.

Original languageEnglish
Pages (from-to)456-464
Number of pages9
JournalWiley Interdisciplinary Reviews: Computational Statistics
Volume5
Issue number6
DOIs
Publication statusPublished - Nov 2013

Fingerprint

Alternating Least Squares
Principal Components
Principal Component Analysis
Optimal Scaling
Nonlinear Analysis
Ordinal Data
Nominal or categorical data
Least Square Algorithm
Alternate
Transform

Keywords

  • Alternating least squares algorithm
  • Homogeneity
  • Optimal scaling
  • Quantification

ASJC Scopus subject areas

  • Statistics and Probability

Cite this

Alternating least squares in nonlinear principal components. / Kuroda, Masahiro; Mori, Yuichi; Iizuka, Masaya; Sakakihara, Michio.

In: Wiley Interdisciplinary Reviews: Computational Statistics, Vol. 5, No. 6, 11.2013, p. 456-464.

Research output: Contribution to journalArticle

Kuroda, Masahiro ; Mori, Yuichi ; Iizuka, Masaya ; Sakakihara, Michio. / Alternating least squares in nonlinear principal components. In: Wiley Interdisciplinary Reviews: Computational Statistics. 2013 ; Vol. 5, No. 6. pp. 456-464.
@article{4584b8ff8edd460e9941aa4579a5ac9a,
title = "Alternating least squares in nonlinear principal components",
abstract = "Principal components analysis (PCA) is probably the most popular descriptive multivariate method for analyzing quantitative data with ratio and interval scale measures. When applying PCA to nominal and ordinal data, the data are processed by a method such as optimal scaling, which nonlinearly transforms nominal and ordinal data into quantitative data. Therefore, PCA with optimal scaling is called nonlinear PCA. Nonlinear PCA reveals nonlinear relationships among variables with different measurement levels and therefore presents a more flexible alternative to ordinary PCA. The alternating least squares algorithm is utilized for nonlinear PCA. The algorithm alternates between optimal scaling for quantifying nominal and ordinal data and ordinary PCA for analyzing optimally scaled data. This article discusses two nonlinear PCA algorithms, namely, PRINCIPALS and PRINCALS.",
keywords = "Alternating least squares algorithm, Homogeneity, Optimal scaling, Quantification",
author = "Masahiro Kuroda and Yuichi Mori and Masaya Iizuka and Michio Sakakihara",
year = "2013",
month = "11",
doi = "10.1002/wics.1279",
language = "English",
volume = "5",
pages = "456--464",
journal = "Wiley Interdisciplinary Reviews: Computational Statistics",
issn = "1939-5108",
publisher = "John Wiley and Sons Inc.",
number = "6",

}

TY - JOUR

T1 - Alternating least squares in nonlinear principal components

AU - Kuroda, Masahiro

AU - Mori, Yuichi

AU - Iizuka, Masaya

AU - Sakakihara, Michio

PY - 2013/11

Y1 - 2013/11

N2 - Principal components analysis (PCA) is probably the most popular descriptive multivariate method for analyzing quantitative data with ratio and interval scale measures. When applying PCA to nominal and ordinal data, the data are processed by a method such as optimal scaling, which nonlinearly transforms nominal and ordinal data into quantitative data. Therefore, PCA with optimal scaling is called nonlinear PCA. Nonlinear PCA reveals nonlinear relationships among variables with different measurement levels and therefore presents a more flexible alternative to ordinary PCA. The alternating least squares algorithm is utilized for nonlinear PCA. The algorithm alternates between optimal scaling for quantifying nominal and ordinal data and ordinary PCA for analyzing optimally scaled data. This article discusses two nonlinear PCA algorithms, namely, PRINCIPALS and PRINCALS.

AB - Principal components analysis (PCA) is probably the most popular descriptive multivariate method for analyzing quantitative data with ratio and interval scale measures. When applying PCA to nominal and ordinal data, the data are processed by a method such as optimal scaling, which nonlinearly transforms nominal and ordinal data into quantitative data. Therefore, PCA with optimal scaling is called nonlinear PCA. Nonlinear PCA reveals nonlinear relationships among variables with different measurement levels and therefore presents a more flexible alternative to ordinary PCA. The alternating least squares algorithm is utilized for nonlinear PCA. The algorithm alternates between optimal scaling for quantifying nominal and ordinal data and ordinary PCA for analyzing optimally scaled data. This article discusses two nonlinear PCA algorithms, namely, PRINCIPALS and PRINCALS.

KW - Alternating least squares algorithm

KW - Homogeneity

KW - Optimal scaling

KW - Quantification

UR - http://www.scopus.com/inward/record.url?scp=84886748803&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84886748803&partnerID=8YFLogxK

U2 - 10.1002/wics.1279

DO - 10.1002/wics.1279

M3 - Article

VL - 5

SP - 456

EP - 464

JO - Wiley Interdisciplinary Reviews: Computational Statistics

JF - Wiley Interdisciplinary Reviews: Computational Statistics

SN - 1939-5108

IS - 6

ER -