📝 Instructions from Markos
Key Requirements:
- 2-13 figures total (maximum 12 for the paper)
- EDS figures BEFORE energy figures (better narrative flow)
- Information-dense figures with good color palettes
- Recreate Markos's 3 screenshots with better quality
- Provide code for both R and Python (Windows-compatible)
🚀 Quick Start
⚠️ Python Not Available?
Use the R script instead! The R script generates identical high-quality figures and doesn't require Python.
🤖 Model Code
XGBoost Models with SHAP
The paper uses XGBoost models with SHAP for explainable machine learning:
- EDS Model: 13 variables, R² = 21.1% (CV)
- Energy Model (with EDS): 14 variables, R² = 82.3% (CV)
- Energy Model (without EDS): 13 variables, R² = 51.6% (CV)
R Scripts for Models (Windows Compatible)
Main Model Script
📥 Download EDS/Energy Models (R)XGBoost models with SHAP for EDS and energy prediction
Model Results
Model results and SHAP values are saved as RDS files in:
paper_a_analysis/data/processed/
- model_results.rds
- shap_values_eds.rds
- shap_values_energy.rds