| """Module: scripts/__init__.py - Scripts Package ALICE. |
| |
| Document ID: ALICE-MOD-SCRIPTS-001 |
| Version: 0.5.0 |
| |
| Scripts utilitaires et pipelines ML pour le moteur ALICE. |
| |
| Packages: |
| - training/: Pipeline d'entraînement ML (CatBoost, XGBoost, LightGBM) |
| - features/: Feature engineering (ALI, CE) |
| - fairness/: Détection de biais (ISO 24027) |
| - robustness/: Tests de robustesse (ISO 24029) |
| - model_registry/: Gestion modèles production |
| - evaluation/: Métriques et benchmarks |
| |
| Scripts principaux: |
| - train_models_parallel.py: Entraînement parallèle multi-modèles |
| - feature_engineering.py: Extraction features ML |
| - ensemble_stacking.py: Stacking ensemble |
| - evaluate_models.py: Benchmarks et évaluation |
| - parse_dataset/: Parsing données FFE |
| |
| ISO Compliance: |
| - ISO/IEC 42001:2023 - AI Management System (pipelines ML) |
| - ISO/IEC 5259:2024 - Data Quality for ML (lineage) |
| - ISO/IEC 5055:2021 - Code Quality (modules <300 lignes) |
| - ISO/IEC TR 24027:2021 - Bias in AI (fairness/) |
| - ISO/IEC 24029:2021 - Neural Network Robustness (robustness/) |
| - ISO/IEC 25059:2023 - AI Quality Model |
| |
| Author: ALICE Engine Team |
| Last Updated: 2026-01-10 |
| """ |
|
|
