Reproduce “Algorithmic Accountability in Small Data: Sample-Size-Induced Bias Within Classification Metrics” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jarren Briscoe, Garrett Kepler, Daryl Deford, Assefaw Gebremedhin
The authors examine how small samples affect classification metrics calculated from confusion matrices. Scores fluctuate because four counts are discrete, and some metrics are undefined when denominators are zero. They propose MATCH to compare scores with reference distributions and CPS to stabilize counts using other groups. Across 15 metrics and two datasets, CPS lowered mean-squared error in downsampling tests, assuming an informative reference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.