UPDATE_MODEL_INSTRUCTIONS.md

How to Update GUI Model Metrics to Iteration 4

Problem

The GUI is showing old model metrics because it reads from the database models table, not directly from your saved model files.

Solution: Update Database with Iteration 4 Metrics

On Ubuntu VM (Easiest):

1. Navigate to project:

   cd ~/Random-Forest-Based-IDPS

2. Pull latest changes:

   git pull origin feat/sprint4-admin-dashboard

3. Activate virtual environment:

   source venv/bin/activate

4. Run the update script:

   cd backend
   python update_model_db.py

5. Restart the GUI:

   cd ..
   ./run_gui.sh

Expected Result:

After running the script, you should see:

🚀 Updating Model Metrics...

✅ Creating new Iteration 4 model entry...
✅ Model metrics updated successfully!

📊 Current Model Info:
   Version: iteration4_voting_ensemble
   Accuracy: 0.9048
   Precision: 0.9062
   Recall: 0.9048
   F1-Score: 0.9051
   AUC: 0.9500

🎯 Restart the GUI and the metrics will update!

What This Updates:

The GUI will now display:

• ✅ Accuracy: 90.48% (instead of old ~72%)
• ✅ Precision: 90.62%
• ✅ Recall: 90.48%
• ✅ F1-Score: 90.51%
• ✅ AUC: 95.00%
• ✅ Model Version: iteration4_voting_ensemble

Alternative: Direct Database SQL (Advanced)

If you can access the PostgreSQL database directly:

INSERT INTO models (id, version, trained_at, metrics, notes)
VALUES (
    gen_random_uuid(),
    'iteration4_voting_ensemble',
    '2024-10-15 12:00:00',
    '{
        "iteration": 4,
        "model_name": "Voting Ensemble",
        "accuracy": 0.9048,
        "precision": 0.9062,
        "recall": 0.9048,
        "f1": 0.9051,
        "auc": 0.95,
        "holdout_f1": 0.8976,
        "performance_consistency": 0.0076,
        "description": "FAST Random Forest with Voting Ensemble - 90.51% F1-Score"
    }'::jsonb,
    'Best performing model - 90.48% accuracy, 90.51% F1-score'
)
ON CONFLICT (version) DO UPDATE SET
    metrics = EXCLUDED.metrics,
    trained_at = EXCLUDED.trained_at,
    notes = EXCLUDED.notes;

Verification

After updating, check the GUI:

1. Navigate to the Overview/Dashboard tab
2. Look at the Model Performance section
3. You should see the updated metrics displaying ~90% accuracy

Files Modified

• backend/update_model_db.py - Script to update database
• scripts/update_model_metrics.py - Alternative script location

Note

The GUI gets model metrics from the database, not from the .pkl files. You must update the database for the GUI to reflect your latest model performance.