Research map / Learning theory and algorithms / Kernels and regression theory
On Mahalanobis Distance in Functional Settings
José R. Berrendero, Beatriz Bueno-Larraz, Antonio Cuevas
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 |
|---|---|---|---|
| increase contamination rate from 0.05 to 0.2 | Mah. RKHS · Outlier detection Model 1 | correct outlier rate p_c: 0.998 -> 0.141 | Hurt |
| increase contamination rate from 0.05 to 0.2 | Mah. RKHS · Outlier detection Model 3 | correct outlier rate p_c: 1.000 -> 0.749 | Hurt |
| increase contamination rate from 0.05 to 0.2 | Mah. RKHS · Outlier detection Model 2 | correct outlier rate p_c: 1.000 -> 0.945 | No clear effect |
| use Mah. RKHS instead of robust Mahalanobis distance | Mah. RKHS · Outlier detection Model 3 (c=0.2) | correct outlier rate p_c: 0.111 -> 0.749 | Improved |
| use Mah. RKHS instead of robust Mahalanobis distance | Mah. RKHS · Outlier detection Model 1 (c=0.2) | correct outlier rate p_c: 0.866 -> 0.141 | Hurt |
| use M_alpha instead of knn5 | M_alpha · Cut Brownian motion vs bridge (T=1) | misclassification %: 27.6 -> 26.5 | Improved |
| use M_alpha instead of knn5 | M_alpha · Cut Brownian motion vs bridge (T=0.75) | misclassification %: 42.4 -> 42.5 | No clear effect |
| use M_alpha instead of OB | M_alpha · Scenario B, n=100, different mean and sd | misclassification %: 30.1 -> 7.7 | Improved |
| use M_alpha instead of OB | M_alpha · Scenario A, n=100, same mean, different sd | misclassification %: 9.3 -> 34.2 | Hurt |
| use M_alpha instead of knn5 | M_alpha · Scenario B, n=50, different mean and same sd | misclassification %: 37.2 -> 41.8 | Hurt |
| increase training size from 50 to 100 | M_alpha · Scenario C, different mean and sd | misclassification %: 19.9 -> 13.3 | Improved |
| choose alpha automatically by reducing KL divergence | Mah. RKHS boxplots · Berkeley growth | alpha selected: 0.089 and 0.1 | Mixed |
Closest papers by meaning
- Kernel Quantile Embeddings and Associated Probability Metrics (ICML 2025 · similarity 0.83)
- An Optimized Franz-Parisi Criterion and its Equivalence with SQ Lower Bounds (NeurIPS 2025 · similarity 0.82)
- A New Robust Partial p-Wasserstein-Based Metric for Comparing Distributions (ICML 2024 · similarity 0.82)
- Hierarchical Integral Probability Metrics: A distance on random probability measures with low sample complexity (ICML 2024 · similarity 0.82)
- The Value of Covariance Matching in Gaussian DDPMs and the Lanczos Sampler (NeurIPS 2026 · similarity 0.82)
- Johnson-Lindenstrauss Lemma Beyond Euclidean Geometry (NeurIPS 2025 · similarity 0.81)
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.
