Research map / Learning theory and algorithms / Prediction and decision losses
A Comparative Evaluation of Quantification Methods
Tobias Schumacher, Markus Strohmaier, Florian Lemmerich
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 |
|---|---|---|---|
| add accuracy-tuned logistic regression base classifier (ablation) | AC · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.332 -> 0.262 | Improved |
| add accuracy-tuned logistic regression base classifier (ablation) | PAC · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.155 -> 0.213 | Hurt |
| add accuracy-tuned logistic regression base classifier (ablation) | TSX · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.151 -> 0.239 | Hurt |
| add accuracy-tuned logistic regression base classifier (ablation) | TSMax · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.140 -> 0.249 | Hurt |
| add accuracy-tuned logistic regression base classifier (ablation) | Median Sweep · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.124 -> 0.236 | Hurt |
| add accuracy-tuned SVM base classifier (ablation) | Median Sweep · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.124 -> 0.188 | Hurt |
| add accuracy-tuned SVM base classifier (ablation) | AC · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.332 -> 0.250 | Improved |
| use random forest base classifier instead of logistic regression (ablation) | AC · Binary quantification, 24 datasets (tuned base classifiers) | absolute error (AE): 0.332 -> 0.259 | Improved |
| use Median Sweep threshold selection instead of plain Adjusted Count | AC (logistic regression) · Binary quantification, 40 datasets | absolute error (AE): 0.324 -> 0.088 | Improved |
| use Median Sweep instead of Classify and Count | CC (logistic regression) · Binary quantification, 40 datasets | absolute error (AE): 0.529 -> 0.088 | Improved |
| use HDx feature-distribution matching instead of Adjusted Count | AC (logistic regression) · Binary quantification, 40 datasets | NKLD: 0.254 -> 0.032 | Improved |
| use quantification-oriented classifiers (SVM-K, SVM-Q, QF) instead of Classify and Count | CC · Binary quantification, 40 datasets | absolute error (AE): 0.529 -> 0.414 | Mixed |
Closest papers by meaning
- DyS: A Framework for Mixture Models in Quantification (AAAI 2019 · similarity 0.90)
- An Equivalence Analysis of Binary Quantification Methods (AAAI 2023 · similarity 0.88)
- quanda: An Interpretability Toolkit for Training Data Attribution Evaluation (NeurIPS 2026 · similarity 0.86)
- SimPro: A Simple Probabilistic Framework Towards Realistic Long-Tailed Semi-Supervised Learning (ICML 2024 · similarity 0.86)
- Evaluating multiple models using labeled and unlabeled data (NeurIPS 2025 · similarity 0.85)
- A Data-Centric Perspective on Evaluating Machine Learning Models for Tabular Data (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.
