Freelance data scientist · All case studies

Mortality and duration, modelled as two questions.

Real breast-cancer clinical profiles, split with care: a classifier for mortality status, ensembles and hypers, and a separate regression for survival duration. One dataset, two honest targets.

Client: Oncology outcomes study. Built by Dilshad Raza.

Clinical tables mix status flags and durations. Stuffing both into one accuracy number hides the harder question. The profiles had to be subsetted properly, leakage watched, and mortality classification kept separate from survival regression.

I split the clinical profiles into the subsets the questions required, trained a mortality classifier with hyperparameter search and ensemble methods, and engineered regression models for survival duration so a “how long” answer did not pretend to be a “whether” answer.

Hire Dilshad Raza for similar freelance data science, machine learning, and AI automation in the UK, United States, and Australia.