Patient Readmission Risk Prediction Model

Construct a sophisticated machine learning model for predicting patient readmission risks using comprehensive healthcare datasets. Develop a Python solution using gradient boosting algorithms that can analyze multiple feature sets including diagnostic codes, treatment history, socioeconomic factors, and patient demographics. Create a Flask-based scoring service that provides real-time readmission probability with interpretable feature importance visualization.