A shrinkage principle for heavy-tailed data: High-dimensional robust low-rank matrix recovery

Mathematics

• Low-rank Matrices 100%
• High-dimensional Data 87%
• Shrinkage 85%
• Recovery 73%
• Moment 39%
• Linear Model 33%
• Multi-task Learning 32%
• Robust Inference 26%
• Matrix Completion 26%
• Compressed Sensing 25%
• Penalized Least Squares 25%
• Truncated Data 25%
• Concentration Inequalities 24%
• Spectral Norm 24%
• Sample Covariance Matrix 23%
• Estimator 22%
• Optimal Rate of Convergence 22%
• Moment Conditions 22%
• Robust Estimation 22%
• Error Rate 21%
• Statistical Inference 18%
• Higher Dimensions 17%
• Optimality 15%
• Regression Model 15%
• High-dimensional 15%
• Trace 15%
• Regression 14%
• Methodology 13%
• Simulation 11%
• Design 11%

Business & Economics

• Shrinkage 99%
• Matrix 67%
• Estimator 30%
• Statistical Inference 26%
• Moment Conditions 25%
• Robust Estimation 25%
• Optimality 23%
• Rate of Convergence 23%
• By-products 22%
• Covariance Matrix 21%
• Least Squares 20%
• Regression Model 15%
• Simulation 12%
• Methodology 9%