Research map / Learning theory and algorithms / Prediction and decision losses
Statistical Comparisons of Classifiers by Generalized Stochastic Dominance
Christoph Jansen, Malte Nalenz, Georg Schollmeyer, Thomas Augustin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use delta=10^-5 instead of delta=0 in the dominance test (ablation) | GSD test · Simulation (12 scenarios) | F-score | Improved |
| raise threshold delta from 0 to 0.004 (ablation) | GSD-delta · UCI 16 data sets | comparable pairs: delta 0 -> 0.004: RIDGE dominates EN and LASSO | Improved |
| raise threshold delta to delta_max (ablation) | GSD-delta · UCI 16 data sets | comparable pairs: delta_max=0.0077: GLM dominates EN, LASSO and RIDGE | Improved |
| treat criteria as ordinal (classical stochastic dominance) instead of using metric information (ablation) | GSD-delta · UCI 16 data sets | comparable pairs: delta=0: only two comparable pairs; delta=0.06 needed, RF and GBM still incomparable | Hurt |
| use GSD test instead of all-test heuristic | GSD / GSD-delta test · Simulation (12 scenarios) | F-score | Improved |
| use GSD test instead of one-test heuristic | GSD / GSD-delta test · Simulation, eta>=0.05 and >=15 data sets | F-score: eta >= 0.05, d >= 15 | Improved |
| drop Bonferroni correction (ablation) | GSD / GSD-delta test · Simulation (12 scenarios) | F-score | Improved |
| increase separation parameter eta and number of data sets | All methods · Simulation (12 scenarios) | F-score | Improved |
| increase resamples from N=1000 to N=10000 (ablation) | Dominance test, delta_min · UCI 16 data sets | significant pairs: N=1000 -> N=10000: still only GBM over BDS significant | No clear effect |
| compare in-sample orders: GSD vs rank-based | GSD-delta · Simulation (12 scenarios) | F-score (sample order) | No clear effect |
Closest papers by meaning
- Statistical Multicriteria Benchmarking via the GSD-Front (NeurIPS 2024 · similarity 0.90)
- Multivariate Stochastic Dominance via Optimal Transport and Applications to Models Benchmarking (NeurIPS 2024 · similarity 0.85)
- Class-Attribute Priors: Adapting Optimization to Heterogeneity and Fairness Objective (AAAI 2024 · similarity 0.85)
- How many classifiers do we need? (NeurIPS 2024 · similarity 0.84)
- Exact power indices for plurality-voting ensembles (NeurIPS 2026 · similarity 0.84)
- Most Influential Subset Selection: Challenges, Promises, and Beyond (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
