Research map / Label-efficient and robust learning / Recognition and long-tail
Multi-Response Linear Discriminant Analysis in High Dimensions
Kai Deng, Xin Zhang, Aaron J. Molstad
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 |
|---|---|---|---|
| fit MLDA jointly across responses instead of DSDA applied marginally to each response | MLDA · Yeast (joint classification) | joint classification error rate (%): 85.0 -> 77.7 | Improved |
| fit MLDA jointly instead of Logistic regression per response | MLDA · Image (joint classification) | joint classification error rate (%): 82.1 -> 51.7 | Improved |
| model response dependence with MLDA instead of pairwise MSDA-P | MLDA · Yeast (joint classification) | joint classification error rate (%): 80.9 -> 77.7 | Improved |
| use full joint likelihood MLDA instead of pairwise likelihood MLDA-P (ablation) | MLDA · Image (joint classification) | joint classification error rate (%): 57.0 -> 51.7 | Improved |
| restrict pairwise MLDA to labels observed in training (MLDAs-P) instead of all combinations (ablation) | MLDA-P · Image (joint classification) | joint classification error rate (%): 57.0 -> 52.1 | Improved |
| restrict pairwise MLDA to labels observed in training (MLDAs-P) instead of all combinations (ablation) | MLDA-P · VirusGO (joint classification) | joint classification error rate (%): 20.2 -> 20.1 | Mixed |
| use MLDA instead of DSDA on binary responses | MLDA · VirusGO (joint classification) | joint classification error rate (%): 17.8 -> 17.8 | No clear effect |
| use MLDA instead of DSDA on binary responses | MLDA · GnegativeGO (joint classification) | joint classification error rate (%): 8.0 -> 7.2 | Improved |
| add conditional refinement (MLDA-C) to MLDA when the other response is given | MLDA · simulation M2 conditional classification | conditional classification error | Improved |
| add conditional refinement (MLDA-C) to MLDA | MLDA · simulation M2 conditional classification | false positive rate of variable selection: 18% -> close to 0% | Improved |
| fit MSDA on partial data (MSDA-S) instead of all data | MSDA · simulation M3 and M4 conditional classification | conditional classification error | Hurt |
| fit MSDA on partial data (MSDA-S) instead of all data | MSDA · simulation M2 and M5 conditional classification | conditional classification error | Improved |
Closest papers by meaning
- High-dimensional Linear Discriminant Analysis Classifier for Spiked Covariance Model (JMLR 2020 · similarity 0.87)
- Latent Discriminant Analysis with Representative Feature Discovery (AAAI 2017 · similarity 0.86)
- A Direct Approach for Sparse Quadratic Discriminant Analysis (JMLR 2018 · similarity 0.86)
- Separability-Oriented Subclass Discriminant Analysis (TPAMI 2017 · similarity 0.86)
- A Generalized Geometric Theoretical Framework of Centroid Discriminant Analysis for Linear Classification of Multi-dimensional Data (ICLR 2026 · similarity 0.85)
- High-Dimensional Tensor Regression With Oracle Properties (ICML 2025 · similarity 0.85)
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.
