This paper introduces a novel approach to transfer learning for linear discriminant analysis in high-dimensional two-class classification. It decomposes the mean difference in each domain into a shared classification signal and a domain-specific deviation. The research derives deterministic limits for classification error under various covariance settings, quantifying the impact of shared signals and domain-specific variations on transfer performance. The findings also lead to improved transfer weights and estimators, along with a correction for intercept bias caused by unbalanced class sample sizes. AI
IMPACT This research could improve the accuracy and efficiency of classification models in scenarios with limited or varied data.
RANK_REASON The cluster contains an academic paper detailing a new methodology in statistics and machine learning.
- alphaXiv
- arXiv
- CatalyzeX Code Finder for Papers
- Connected Papers
- CORE Recommender
- DagsHub
- Gotit.pub
- Hugging Face
- linear discriminant analysis
- Litmaps
- ScienceCast
- scite Smart Citations
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
