Principal component analysis for compositional data with outliers

By Peter Filzmoser, Karel Hron, Clemens Reimann. 2009

535 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Mechanical Engineering, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/36c481e6fd532ebb0bbd5f1da41386147f652fd8

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Matthias Templ
2. Karel Hron
3. Peter Filzmoser
4. J. Palarea‐Albaladejo
5. Eric C. Grunsky
6. Josep-Antoni Martín-Fernández
7. Eva Fiserová
8. Janice L. Scealy

In Mechanical Engineering

1. Eric C. Grunsky
2. Matthias Templ
3. Karel Hron
4. Eva Fiserová
5. Alan H. Welsh
6. Janice L. Scealy
7. Josep-Antoni Martín-Fernández
8. J. Palarea‐Albaladejo

In Statistics and Probability

1. Matthias Templ
2. Karel Hron
3. Eva Fiserová
4. Peter Filzmoser
5. J. Palarea‐Albaladejo
6. Josep-Antoni Martín-Fernández
7. Eric C. Grunsky
8. Ricardo A. Olea

Citations in the corpus

• Dealing with Zeros (2011, methodology): “Note that, due to the high sample variability observed in the data, the principal components from the biplot were obtained using robust methods (Filzmoser et al., 2009).”
• robCompositions: An R‐package for Robust Statistical Analysis of Compositional Data (2011, methodology): “…(Templ et al., 2009) contains functions for robust statistical methods designed for compositional data, like principal component analysis (Filzmoser et al., 2009a) (including the robust compositional biplot), factor analysis (Filzmoser et al., 2009b), and discriminant analysis (Filzmoser…”
• Robust Principal Component Analysis for Power Transformed Compositional Data (2015, methodology): “…Section 3.3 with α = 0 (GR); Euclidean PCA applied to the centered log-ratio transformed data (CLR); the robust spherical PCA method of Locantore et al. (1999) applied to the centered log-ratio transformed data (LOC); and the robust method of Filzmoser et al. (2009) described in Section 3.1 (MCD).”
• On the Interpretation of Orthonormal Coordinates for Compositional Data (2011, methodology): “This is done in order to construct the compositional biplot (Aitchison and Greenacre 2002; Filzmoser et al. 2009a ), for instance, or even to perform some other statistical methods such as factor analysis (Filzmoser et al. 2009c).”
• Imputation of missing values for compositional data using classical and robust methods (2010, methodology): “Applying standard statistical methods like correlation analysis or principal component analysis directly to compositional data would give misleading results (Pearson, 1897; Aitchison, 1986; Filzmoser and Hron, 2009; Filzmoser et al., 2009).”