How elazzouzi1080/tracia_usecases works
This repository provides a reproducible binary-classification experiment workflow: prepare stable stratified folds, run baseline models against those folds, evaluate their predictions, and export summarized results. The README documents visualizations, confidence intervals, statistical comparisons, and CSV/JSON/Excel outputs; the sampled Python excerpt confirms the loading, evaluation, orchestration, and saving layers. Details of unsampled notebook cells and analysis methods are not asserted beyond the documented capabilities.
Main entry points: fixed_cv_binary_classification.py Pipeline orchestrator · 01_prepare_splits_and_config.ipynb Prepare splits · 02_run_binary_classification_pipeline.ipynb Run evaluation · splits_k5_v1 Fixed fold data
Components 17
Data preparation
- Prepare splits
use_cases/notebooks_usecases/01_prepare_splits_and_config.ipynb - Clinical dataset
use_cases/data/heart_failure_clinical_records_dataset.csv - Fixed fold data
use_cases/data/splits_k5_v1
Model evaluation
- Run evaluation
use_cases/notebooks_usecases/02_run_binary_classification_pipeline.ipynb - Pipeline orchestrator
use_cases/notebooks_usecases/fixed_cv_binary_classification.py - Cross-validation evaluator
use_cases/notebooks_usecases/fixed_cv_binary_classification.py - Data loader
use_cases/notebooks_usecases/fixed_cv_binary_classification.py - Baseline models
use_cases/notebooks_usecases/fixed_cv_binary_classification.py - Metric calculation
use_cases/notebooks_usecases/fixed_cv_binary_classification.py
Results analysis
- Results analysis
use_cases/notebooks_usecases/fixed_cv_binary_classification.py - Evaluation figures
use_cases/results_pipline/figures
Experiment artifacts
- Per-fold results
use_cases/results_pipline/cv_results_per_fold.csv - Summary results
use_cases/results_pipline/cv_results_summary.csv - JSON exports
use_cases/results_pipline/cv_results_per_fold.json - Results workbook
use_cases/results_pipline/cv_results_complete.xlsx
Other components
- Experimenter
- Scikit-learn
Connections 17
- Experimenter → Prepare splits: prepares folds
- Prepare splits → Clinical dataset: reads dataset
- Prepare splits → Fixed fold data: writes folds
- Experimenter → Run evaluation: starts evaluation
- Run evaluation → Pipeline orchestrator: runs pipeline
- Pipeline orchestrator → Cross-validation evaluator: runs evaluation
- Pipeline orchestrator → Results analysis: summarizes results
- Cross-validation evaluator → Data loader: loads data
- Cross-validation evaluator → Baseline models: creates models
- Cross-validation evaluator → Metric calculation: calculates metrics
- Data loader → Fixed fold data: reads dataset and folds
- Baseline models → Scikit-learn: uses estimators
- Pipeline orchestrator → Per-fold results: writes CSV
- Pipeline orchestrator → Summary results: writes CSV
- Pipeline orchestrator → JSON exports: writes JSON
- Pipeline orchestrator → Results workbook: writes workbook
- Results analysis → Evaluation figures: creates figures