(積ん読)Approximate statistical tests for comparing supervised classification learning algorithms
https://sci2s.ugr.es/keel/pdf/algorithm/articulo/dietterich1998.pdf
https://direct.mit.edu/neco/article-abstract/10/7/1895/6224/Approximate-Statistical-Tests-for-Comparing?redirectedFrom=PDF
Model Evaluation, Model Selection, and Algorithm Selection in Machine Learning
と同様の切り分け方で議論しているらしい(
4.8 Comparing Algorithms
)
McNemar testが好ましいという結論らしい(
4.3 Comparing Two Models with the McNemar Test
)