This repository contains the materials and code for my Bayesian statistics coursework as part of the MPhil in Health Data Science program, University of Cambridge. The coursework focuses on analyzing patient-specific biomarker data using Bayesian hierarchical modeling.
The pdf file (Bayesian_coursework.pdf) contains my full analysis.
The study examines data from 12 patients administered a drug (SEYAB) and tracks the levels of a biomarker (SGAJ) at regular intervals. The primary tasks include:
- Visualizing Patient Data: Explore the influence of age and sex on SGAJ levels over time.
- Bayesian Hierarchical Modeling: Build a hierarchical model to estimate the rate of change in SGAJ levels and assess inter-patient variability.
- Prediction: Predict biomarker levels at a future time point and evaluate the probability of exceeding a threshold.
- Model Validation: Conduct posterior predictive checks to assess model fit.
- Hierarchical Bayesian modeling to account for variability across patients.
- Use of prior knowledge to inform model assumptions.
- Posterior predictive checks to validate model performance.
- Insights into the relationship between patient characteristics (age, sex) and biomarker dynamics.