Publication type: | Conference paper |
Type of review: | Peer review (publication) |
Title: | Can we ignore the compositional nature of compositional data by using deep learning aproaches? |
Authors: | Templ, Matthias |
et. al: | No |
Proceedings: | Book of Short Papers SIS 2021 |
Editors of the parent work: | Perna, Cirna Salvati, Nicola Schirripa Spagnolo, Francesco |
Page(s): | 243 |
Pages to: | 248 |
Conference details: | Surface Inspection Summit Europe, Aachen, Germany, 9-10 November 2021 |
Issue Date: | 2021 |
Publisher / Ed. Institution: | Pearson |
Publisher / Ed. Institution: | London |
ISBN: | 9788891927361 |
Language: | English |
Subjects: | Deep learning; Compositional data analysis |
Subject (DDC): | 006: Special computer methods |
Abstract: | Care must be taken not to simply apply multivariate data analysis methods to compositional data. For example, one can show that correlations are biased to be negative, and almost all statistical methods result in biased estimates when applied to compositional data. One way out is to analyze data methods from compositional data analysis, i.e. by carrying out a log-ratio analysis. This contribution has its focus on settings where only the prediction and classification error is important rather than an interpretation of results. In this setting it is well-known that classification and prediction errors are smaller with a log-ratio approach using traditional machine learning methods. However, is this also true when training a neural network who may learn the inner relationships between parts of a whole also without representing the data in log-ratios? This contribution give an indication on this matter using one real data set from chemical measurements on beers. |
URI: | https://it.pearson.com/content/dam/region-core/italy/pearson-italy/pdf/Docenti/Universit%C3%A0/pearson-sis-book-2021-parte-1.pdf https://digitalcollection.zhaw.ch/handle/11475/24598 |
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. (2021). Can we ignore the compositional nature of compositional data by using deep learning aproaches? [Conference paper]. In C. Perna, N. Salvati, & F. Schirripa Spagnolo (Eds.), Book of Short Papers SIS 2021 (pp. 243–248). Pearson. https://it.pearson.com/content/dam/region-core/italy/pearson-italy/pdf/Docenti/Universit%C3%A0/pearson-sis-book-2021-parte-1.pdf
Templ, M. (2021) ‘Can we ignore the compositional nature of compositional data by using deep learning aproaches?’, in C. Perna, N. Salvati, and F. Schirripa Spagnolo (eds) Book of Short Papers SIS 2021. London: Pearson, pp. 243–248. Available at: https://it.pearson.com/content/dam/region-core/italy/pearson-italy/pdf/Docenti/Universit%C3%A0/pearson-sis-book-2021-parte-1.pdf.
M. Templ, “Can we ignore the compositional nature of compositional data by using deep learning aproaches?,” in Book of Short Papers SIS 2021, 2021, pp. 243–248. [Online]. Available: https://it.pearson.com/content/dam/region-core/italy/pearson-italy/pdf/Docenti/Universit%C3%A0/pearson-sis-book-2021-parte-1.pdf
TEMPL, Matthias, 2021. Can we ignore the compositional nature of compositional data by using deep learning aproaches? In: Cirna PERNA, Nicola SALVATI und Francesco SCHIRRIPA SPAGNOLO (Hrsg.), Book of Short Papers SIS 2021 [online]. Conference paper. London: Pearson. 2021. S. 243–248. ISBN 9788891927361. Verfügbar unter: https://it.pearson.com/content/dam/region-core/italy/pearson-italy/pdf/Docenti/Universit%C3%A0/pearson-sis-book-2021-parte-1.pdf
Templ, Matthias. 2021. “Can We Ignore the Compositional Nature of Compositional Data by Using Deep Learning Aproaches?” Conference paper. In Book of Short Papers SIS 2021, edited by Cirna Perna, Nicola Salvati, and Francesco Schirripa Spagnolo, 243–48. London: Pearson. https://it.pearson.com/content/dam/region-core/italy/pearson-italy/pdf/Docenti/Universit%C3%A0/pearson-sis-book-2021-parte-1.pdf.
Templ, Matthias. “Can We Ignore the Compositional Nature of Compositional Data by Using Deep Learning Aproaches?” Book of Short Papers SIS 2021, edited by Cirna Perna et al., Pearson, 2021, pp. 243–48, https://it.pearson.com/content/dam/region-core/italy/pearson-italy/pdf/Docenti/Universit%C3%A0/pearson-sis-book-2021-parte-1.pdf.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.