App 15 — ML Pipeline Builder · MₓCᵧ
46 transition-metal chalcogenides · build, train, and interpret an ML pipeline step by step
🗄️
1 · Data
⚗️
2 · Features
🤖
3 · Model
📊
4 · Evaluate
🔍
5 · Interpret
Dataset — MₓCᵧ Transition-Metal Chalcogenides
Filter: Metal
All metals
Fe
Ni
Co
Mn
Cr
Ti
Filter: Chalcogen
All (S,Se,Te)
S
Se
Te
Target
Band Gap (eV) — Regression
Magnetic Ordering — Classification
🔍 Filter
Next: Features →
Band Gap Distribution
Magnetic Ordering Count
Select Features
✓ Select all
✗ Clear
Next: Model →
Feature Correlation Matrix
Model Configuration
Algorithm
Random Forest
Ridge / Logistic
SVM (RBF kernel)
Gradient Boosting
CV folds
3-fold
5-fold
10-fold (LOO~)
n_estimators / C
100
▶ Run CV
Next: Evaluate →
Cross-Validation Results — All Models Compared
Model
CV Score
Performance
Std
Train/Test Evaluation
Test size
20%
30%
▶ Evaluate
Next: Interpret →
Parity Plot
Residual Distribution
Feature Importance (MDI / |Coefficient|)
🔍 Compute Importance
Pipeline Summary
Run all pipeline steps to see the summary.
Predict a New Compound
⚡ Predict