Research map / Label-efficient and robust learning / Recognition and long-tail
A Robust-Equitable Measure for Feature Ranking and Selection
A. Adam Ding, Jennifer G. Dy, Yi Li, Yale Chang
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 RCD instead of MI for ranking noisy features across sample sizes | RCD · Synthetic ranking X3 vs X4 | feature rank correctness: MI X4 0.5 portion rises 1.82 -> 2.97 with n 300 -> 10k while RCD stays 0.52 -> 0.52 | Improved |
| use self-equitable measures (MI, CD2, RCD) instead of Pearson or kernel measures | RCD · Synthetic ranking X1 vs X2 | feature ranking | Improved |
| use RCD instead of CD2 with unequal sample sizes | RCD · Synthetic ranking (X3 n=300) | feature ranking: CD2 X3 3.75 vs X4 44.15 at n=300 vs 10k | Improved |
| use RCD instead of MI and CD2 for KEGG unequal sample sizes | RCD · KEGG metabolic network | feature ranking | Improved |
| increase sample size from 1000 to 20000 (ablation) | CD2 · KEGG metabolic network | ranking of X2 vs X3: X2 over X3 -> X3 over X2 | Hurt |
| increase sample size from 1000 to 20000 (ablation) | RCD · KEGG metabolic network | ranking stability | No clear effect |
| use RCD with mRMR instead of MI or CD2 | RCD · Synthetic additive regression, mRMR (n=1000) | cross-validated MSE | Improved |
| use RCD instead of Pearson on stock data | RCD · Stock data X1 vs X2 | feature selection: RCD 0.68 vs 0.67 (X1 vs X2) | Improved |
Closest papers by meaning
- Measuring Dependence Powerfully and Equitably (JMLR 2016 · similarity 0.87)
- Kernel Feature Selection via Conditional Covariance Minimization (NeurIPS 2017 · similarity 0.86)
- On the Stability of Feature Selection Algorithms (JMLR 2018 · similarity 0.85)
- A Rigorous Information-Theoretic Definition of Redundancy and Relevancy in Feature Selection Based on (Partial) Information Decomposition (JMLR 2023 · similarity 0.85)
- Kernel Partial Correlation Coefficient --- a Measure of Conditional Dependence (JMLR 2022 · similarity 0.84)
- Most Influential Subset Selection: Challenges, Promises, and Beyond (NeurIPS 2024 · similarity 0.83)
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.
