Feature Importance in Gradient Boosting Trees with Cross-Validation Feature Selection

Mathematics

• Boosting 100%
• Feature Selection 99%
• Cross-validation 83%
• Gradient 56%
• Biased 35%
• Demonstrate 33%
• Prediction 29%
• Decision tree 23%
• Categorical variable 23%
• Performance 22%
• Framework 21%
• Computational Cost 18%
• Cardinality 14%

Engineering & Materials Science

• Feature extraction 57%
• Defects 50%
• Decision trees 36%
• Costs 14%

Physics & Astronomy

• gradients 48%
• defects 20%
• predictions 19%
• fixing 19%
• performance 16%
• costs 14%