Machine Learning Workflow with edaflow

This guide consolidates all steps for a complete ML workflow using edaflow, from experiment setup to model deployment.

Step 1: Setup ML Experiment

experiment = ml.setup_ml_experiment(df, 'target')  # DataFrame style
# OR
experiment = ml.setup_ml_experiment(X=X, y=y, val_size=0.15)  # sklearn style

Step 2: Compare Multiple Models

models = {
    'RandomForest': RandomForestClassifier(),
    'LogisticRegression': LogisticRegression()
}
results = ml.compare_models(models, **experiment)

Step 3: Optimize Hyperparameters

model_name = 'LogisticRegression'  # or 'RandomForest' or 'GradientBoosting'

if model_name == 'RandomForest':
    param_distributions = {
        'n_estimators': [100, 200, 300],
        'max_depth': [5, 10, 15, None],
        'min_samples_split': [2, 5, 10]
    }
    model = RandomForestClassifier()
    method = 'grid'
elif model_name == 'GradientBoosting':
    param_distributions = {
        'n_estimators': (50, 200),
        'learning_rate': (0.01, 0.3),
        'max_depth': (3, 8)
    }
    from sklearn.ensemble import GradientBoostingClassifier
    model = GradientBoostingClassifier()
    method = 'bayesian'
elif model_name == 'LogisticRegression':
    param_distributions = {
        'C': [0.01, 0.1, 1, 10, 100],
        'penalty': ['l1', 'l2', 'elasticnet', 'none'],
        'solver': ['lbfgs', 'liblinear', 'saga']
    }
    model = LogisticRegression(max_iter=1000)
    method = 'grid'
else:
    raise ValueError(f"Unknown model_name: {model_name}")

results = ml.optimize_hyperparameters(
    model,
    param_distributions=param_distributions,
    **experiment
)

Step 4: Rank and Select Best Model

best_model_name = ml.rank_models(results, 'accuracy', return_format='list')[0]['model_name']
ranked_df = ml.rank_models(results, 'accuracy')
best_model_traditional = ranked_df.iloc[0]['model']

Step 5: Save Model Artifacts

ml.save_model_artifacts(
    model=results['best_model'],
    model_name=best_model_name,
    experiment_config=experiment,
    performance_metrics=results['cv_results']
)

Step 6: Visualize Learning Curves

ml.plot_learning_curves(results['best_model'], **experiment)

Tips: - All steps above are copy-paste safe and work for RandomForest, LogisticRegression, and GradientBoosting. - For more advanced workflows, see the User Guide and API Reference.