NesPrInDT: Nested undersampling in PrInDT
Applications
2021-08-31 v2
Abstract
In this paper, we extend our PrInDT method (Weihs, Buschfeld 2021) towards additional undersampling of one of the predictors. This helps us to handle multiple unbalanced data sets, i.e. data sets that are not only unbalanced with respect to the class variable but also in one of the predictor variables. Beyond the advantages of such an approach, our study reveals that the balanced accuracy in the full data set can be much lower than in the predictor undersamples. We discuss potential reasons for this problem and draw methodological conclusions for linguistic studies.
Cite
@article{arxiv.2103.14931,
title = {NesPrInDT: Nested undersampling in PrInDT},
author = {Claus Weihs and Sarah Buschfeld},
journal= {arXiv preprint arXiv:2103.14931},
year = {2021}
}
Comments
12 pages, 3 figures