eat(sprint1): add patient risk scoring & anomalies notebook

[Tanvir-ctrl1]

Description

Adds Sprint-1 notebook for patient risk scoring & anomaly detection.

Path

• AI Guardian/Patient_Risk_Scoring_Anomalies.ipynb

Summary

• Computes risk scores from vitals using documented SAFE ranges + WEIGHTS.
• EWMA smoothing and robust z-score thresholding for anomaly flags.
• Parameterized (lookback windows, forecast span); produces timelines/plots.
• Uses dataset “New AI spreadsheet – Sheet1.csv” (can change via DATA_PATH at top).

Todos

• Tested locally on the provided CSV
• Self-reviewed; added Markdown notes in notebook
• Request review from 2 team members

How to test

1. Checkout this PR’s branch.
2. Open AI Guardian/Patient_Risk_Scoring_Anomalies.ipynb in VS Code/Jupyter.
3. If your CSV path differs, update DATA_PATH.
4. Run all cells. Expect: risk timeline plot + anomaly flags without errors.

Screenshots/GIFs

N/A (notebook renders plots inline).

Associated Tasks

N/A

Known Issues

• Notebook assumes the CSV is available; update DATA_PATH if stored elsewhere.