FeaFiner: Biomarker identification from medical data through feature generalization and selection

Keyphrases

• Biomarker Identification 100%
• Medical Data 100%
• Feature Selection 100%
• Feature Generalization 100%
• Low-level Features 66%
• Proposed Formulation 66%
• High-level Concepts 66%
• Clinical Research 33%
• Prediction Accuracy 33%
• Predictive Modeling 33%
• Optimization Problem 33%
• Numerical Experiments 33%
• Refinery 33%
• Learning Process 33%
• Disease Status 33%
• Mini-Mental State Examination 33%
• Mutually Exclusive 33%
• Non-overlapping 33%
• Asymptotic Behavior 33%
• Modeling Method 33%
• Real-world Application 33%
• Computational Efficiency 33%
• Generalization Bounds 33%
• Concept Base 33%
• Augmented Lagrangian Method 33%
• Non-smoothness 33%
• Alzheimer's Disease Neuroimaging Initiative 33%
• Alzheimer's Disease Assessment Scale 33%
• Non-monotone 33%
• Non-convexity 33%
• Feature Grouping 33%
• Double Sparsity 33%
• High-level Features 33%
• Efficient Formulation 33%
• Robust Features 33%
• Feature Construction 33%
• Orthogonality Constraint 33%
• Smoothness Penalty 33%
• Sparsity Optimization 33%
• Spectral Projected Gradient Method 33%
• Selective Features 33%

Computer Science

• Feature Selection 100%
• Feature Extraction 100%
• Data Mining 50%
• Sparsity 50%
• Optimization Problem 50%
• Learning Process 50%
• Interpretability 50%
• World Application 50%
• Gradient Method 50%
• Predictive Accuracy 50%
• Feature Construction 50%
• Separate Process 50%

Biochemistry, Genetics and Molecular Biology

• Feature Extraction 100%
• Biomarker Identification 100%
• Clinical Research 50%
• Mental Health 50%
• Asymptotic Behavior 50%