Survival Analysis Thesis

Context: M.Sc. Statistics thesis – primary field data collection on real orthopedic disease patients.

Challenge: Identify the optimal predictive model for disease recurrence from a set of 6 competing survival models.

Methodology: Conducted end‑to‑end survival analysis across two successive research projects:

• Applied Kaplan‑Meier curves and Log‑Rank tests for initial survival estimation
• Built hazard models and Random Survival Forests for advanced prediction
• Evaluated 6 models: Cox PH, XGBoost, CoxBoost, GLMBoost, GLM, CForest

Outcome: Delivered a comparative evaluation that identified the most accurate model for predicting disease recurrence, with findings documented in a full research thesis.