xpectra.config
Configuration for the xpectra weathered-unknowns study.
Classes
|
- class xpectra.config.TrustConfig(routes: 'tuple[str, ...]'=('norm', 'deriv1', 'deriv2'), model_names: 'tuple[str, ...]'=('Logistic Regression', 'SVM (RBF)', 'Random Forest (200)', 'XGBoost (100)', 'MLP (128, 64)'), calibration_methods: 'tuple[str, ...]'=('none', 'sigmoid', 'isotonic', 'temperature'), cv_folds: 'int' = 5, test_size: 'float' = 0.2, ece_bins: 'int' = 15, dedupe_spectra: 'bool' = True, pca_components_ood: 'int' = 20, knn_k: 'int' = 5, target_coverage: 'float' = 0.9, pca_components_ssl: 'int' = 50, self_train_threshold: 'float' = 0.9, self_train_max_iter: 'int' = 3, label_spread_neighbors: 'int' = 10, carbonyl_band: 'tuple[float, float]'=(1650.0, 1800.0), reference_band: 'tuple[float, float]'=(1420.0, 1470.0), random_state: 'int' = 42, processed_dir: 'Path' = <factory>, results_dir: 'Path' = <factory>, figures_dir: 'Path' = <factory>, cache_dir: 'Path' = <factory>)[source]
Bases:
object- Parameters:
cv_folds (int)
test_size (float)
ece_bins (int)
dedupe_spectra (bool)
pca_components_ood (int)
knn_k (int)
target_coverage (float)
pca_components_ssl (int)
self_train_threshold (float)
self_train_max_iter (int)
label_spread_neighbors (int)
random_state (int)
processed_dir (Path)
results_dir (Path)
figures_dir (Path)
cache_dir (Path)