Exploring outliers in compositional data with structural zeros

Templ, Matthias; Hron, Karel; Filzmoser, Peter

Publication type:	Conference other
Type of review:	Not specified
Title:	Exploring outliers in compositional data with structural zeros
Authors:	Templ, Matthias Hron, Karel Filzmoser, Peter
Conference details:	CoDaWork 2017: 7th International Workshop on Compositional Data Analysis, Siena, Italy, 5-9 June 2017
Issue Date:	2017
Language:	English
Subjects:	Compositional data analysis
Subject (DDC):	510: Mathematics
Abstract:	The analysis of compositional data using the log-ratio approach is based on ratios between compositional parts. Zeros in the parts thus cause serious difficulties for the analysis. This is a particular problem in presence of structural zeros, resulting from a structural process rather than from imprecision of a measurement device. Examples of structural zeros in compositional data are, e.g. expenditures in tobacco for non-smokers, tax payments on shares for non-share holders, time budget on sports for people who do not do any sports at all, etc. Therefore, structural zeros cannot be simply replaced by a non-zero value as it is done, e.g. for values below detection limit or missing values. Instead, zeros have to be incorporated into further statistical processing. We lay the focus on exploratory tools for identifying outliers in compositional data sets with structural zeros. For this purpose, robust Mahalanobis distances are estimated; they are computed either directly for subcompositions determined by their zero patterns or by using imputation as an auxiliary step to improve the efficiency of the estimates. Consequently, we proceed to the subcompositional and subgroup level. For this approach, new theory is formulated that allows to estimate covariances for imputed compositional data and to apply estimations on subgroups using blocks of this covariance matrix (Templ, Hron and Filzmoser, 2016). Moreover, the zero pattern structure is analyzed using PCA for binary data (de Leeuw, 2016) to achieve a comprehensive view of the overall multivariate structure of zeros. The proposed tools are applied to large-scale data from official statistics, where the need for an appropriate treatment of zeros is obvious.
URI:	http://www.compositionaldata.com/codawork2017/abstracts/8thThursday/Oral/ZER-2.pdf https://digitalcollection.zhaw.ch/handle/11475/6034
Fulltext version:	Published version
License (according to publishing contract):	Licence according to publishing contract
Departement:	School of Engineering
Organisational Unit:	Institute of Data Analysis and Process Design (IDP)
Appears in collections:	Publikationen School of Engineering

Files in This Item:

There are no files associated with this item.

Show full item record

Templ, M., Hron, K., & Filzmoser, P. (2017). Exploring outliers in compositional data with structural zeros. CoDaWork 2017: 7th International Workshop on Compositional Data Analysis, Siena, Italy, 5-9 June 2017. http://www.compositionaldata.com/codawork2017/abstracts/8thThursday/Oral/ZER-2.pdf

Templ, M., Hron, K. and Filzmoser, P. (2017) ‘Exploring outliers in compositional data with structural zeros’, in CoDaWork 2017: 7th International Workshop on Compositional Data Analysis, Siena, Italy, 5-9 June 2017. Available at: http://www.compositionaldata.com/codawork2017/abstracts/8thThursday/Oral/ZER-2.pdf.

M. Templ, K. Hron, and P. Filzmoser, “Exploring outliers in compositional data with structural zeros,” in CoDaWork 2017: 7th International Workshop on Compositional Data Analysis, Siena, Italy, 5-9 June 2017, 2017. [Online]. Available: http://www.compositionaldata.com/codawork2017/abstracts/8thThursday/Oral/ZER-2.pdf

TEMPL, Matthias, Karel HRON und Peter FILZMOSER, 2017. Exploring outliers in compositional data with structural zeros. In: CoDaWork 2017: 7th International Workshop on Compositional Data Analysis, Siena, Italy, 5-9 June 2017 [online]. Conference presentation. 2017. Verfügbar unter: http://www.compositionaldata.com/codawork2017/abstracts/8thThursday/Oral/ZER-2.pdf

Templ, Matthias, Karel Hron, and Peter Filzmoser. 2017. “Exploring Outliers in Compositional Data with Structural Zeros.” Conference presentation. In CoDaWork 2017: 7th International Workshop on Compositional Data Analysis, Siena, Italy, 5-9 June 2017. http://www.compositionaldata.com/codawork2017/abstracts/8thThursday/Oral/ZER-2.pdf.

Templ, Matthias, et al. “Exploring Outliers in Compositional Data with Structural Zeros.” CoDaWork 2017: 7th International Workshop on Compositional Data Analysis, Siena, Italy, 5-9 June 2017, 2017, http://www.compositionaldata.com/codawork2017/abstracts/8thThursday/Oral/ZER-2.pdf.