Sparse Classification with Positive-Confidence Data in High Dimensions
Research Paper#Machine Learning, High-Dimensional Statistics, Sparse Learning, Weak Supervision🔬 Research|Analyzed: Jan 3, 2026 17:12•
Published: Dec 30, 2025 19:53
•1 min read
•ArXivAnalysis
This paper addresses the challenge of high-dimensional classification when only positive samples with confidence scores are available (Positive-Confidence or Pconf learning). It proposes a novel sparse-penalization framework using Lasso, SCAD, and MCP penalties to improve prediction and variable selection in this weak-supervision setting. The paper provides theoretical guarantees and an efficient algorithm, demonstrating performance comparable to fully supervised methods.
Key Takeaways
- •Addresses high-dimensional classification with Positive-Confidence data.
- •Proposes sparse-penalization methods using Lasso, SCAD, and MCP.
- •Provides theoretical guarantees (estimation and prediction error bounds).
- •Develops an efficient proximal gradient algorithm.
- •Achieves performance comparable to fully supervised approaches.
Reference / Citation
View Original"The paper proposes a novel sparse-penalization framework for high-dimensional Pconf classification."