Dataset for EVA 2023 Data Challenge
This data set provides the datasets generated by the three creators (data challenge organisers) and subsequently provided to the participants of the EVA 2023 Data Challenge.
The dataset aims to capture the variety of contexts experienced in the analysis of environmental extremes data. This involves both univariate and multivariate problems. The univariate extremes problems involve inference for extreme quantiles when faced with additional complications such as covariates; data missing at random; and the need to convert the inference into design levels which account for different losses from over- and under-design.
The data set consists of five data files:
1. Amaurot: Training data given to the participants for Tasks 1 and 2
2. AmaurotTestSet: Collection of test data points for which predictions had to be submitted
3. Coputopia: Data participants had to consider for Task 3
4. UtopulaU1 + UtopulaU2: Data participants had to consider for Task 4
The aim of this dataset, developed for the Data Challenge, is to assess performance in multivariate extremes in a way that is independent of marginal extremes abilities. Consequently, the multivariate problems relate to data where the univariate marginal distributions are all known.
Cite this dataset as:
Rohrbeck, C.,
Simpson, E.,
Tawn, J.,
2025.
Dataset for EVA 2023 Data Challenge.
Bath: University of Bath Research Data Archive.
Available from: https://doi.org/10.15125/BATH-01399.
Export
Data
EVA 2023 Challenge Data.zip
application/zip (2MB)
Creative Commons: Attribution 4.0
These data sets were generated in R for the Data Challenge organised as part of the Extreme Value Analysis 2023 conference.
Creators
Christian Rohrbeck
University of Bath
Emma Simpson
University College London
Jonathan Tawn
Lancaster University
Contributors
University of Bath
Hosting Institution
Documentation
Data collection method:
Data is entirely simulated using methodology from the statistical research area of Extreme Value Theory. Further details on the methodology can be found in the associated paper.
Technical details and requirements:
Data was simulated using the statistical programming language R.
Funders
University of Bath
https://doi.org/10.13039/501100000835
Publication details
Publication date: 29 March 2025
by: University of Bath
Version: 1
DOI: https://doi.org/10.15125/BATH-01399
URL for this record: https://researchdata.bath.ac.uk/1399
Related papers and books
Contact information
Please contact the Research Data Service in the first instance for all matters concerning this item.
Contact person: Christian Rohrbeck
Faculty of Science
Mathematical Sciences