Course modules

Machine Learning in Biostatistics Modules

Explore the full five-module structure of the course, from core foundations to applied medical machine learning case studies. Module 1 is complete and ready to study.

Course route

Start with prediction thinking before modelling.

The sequence moves from foundational interpretation discipline to supervised models, validation, modern prediction methods and applied clinical case studies.

Open Module 1 Start Lesson 1.1 View all modules

Modules

Lessons

Available now

Complete module

Module 1 is complete

All five foundation lessons are available with lectures, detailed notes, interactive labs, R labs, reports and quizzes.

25-lesson structure

The course is organised into five modules, each containing five carefully sequenced lessons.

Clinical ML focus

The modules move from prediction thinking to supervised models, validation, modern ML and applied case studies.

Full course structure

Five modules, each with five focused lessons.

The course is designed as a complete pathway. Students first learn how to think about prediction responsibly, then move into methods, validation, modern models and applied reporting.

Module

Complete5 lessons

Foundations of Machine Learning in Biostatistics

Build the core language of prediction, explanation, causality, learning types, train/test validation, overfitting, leakage, thresholds and responsible reporting.

1.1 What is machine learning in biostatistics?Open 1.2 Prediction vs explanation vs causal thinkingOpen 1.3 Types of learningOpen 1.4 Training, testing, overfitting and generalisationOpen 1.5 Biostatistical ML workflowOpen

Open module →

Module

Preparing5 lessons

Supervised Learning for Clinical and Health Data

Move from foundations into supervised prediction models for health data: regression, logistic classification, k-nearest neighbours, trees and clinical modelling pipelines.

2.1 Regression as a prediction modelReady 2.2 Logistic regression as a classifierReady 2.3 k-nearest neighbours and distance-based learningReady 2.4 Decision trees and rule-based predictionReady 2.5 Model pipelines for clinical datasetsReady

Open module →

Module

Preparing5 lessons

Model Evaluation, Validation and Performance

Study the central performance tools for medical prediction: resampling, cross-validation, bootstrap validation, ROC/AUC, calibration and clinical usefulness.

3.1 Train/test split and resamplingReady 3.2 Cross-validation and bootstrap validationReady 3.3 Classification metrics: sensitivity, specificity, ROC, AUCReady 3.4 Calibration, clinical usefulness and decision curvesReady 3.5 Bias, leakage and reproducibility in health MLReady

Open module →

Module

Preparing5 lessons

Regularisation, Ensembles and Modern Prediction Models

Learn flexible prediction methods and how to compare them responsibly: ridge, lasso, elastic net, random forests, boosting and support vector machines.

4.1 Ridge, lasso and elastic netReady 4.2 Random forestsReady 4.3 Gradient boostingReady 4.4 Support vector machines and flexible boundariesReady 4.5 Comparing models responsiblyReady

Open module →

Module

Preparing5 lessons

Applied Biostatistical ML Case Studies

Apply the full workflow to realistic health-data projects: clinical risk prediction, survival outcomes, omics, missing data, imbalance, fairness and final reporting.

5.1 Clinical risk prediction case studyReady 5.2 Survival prediction and censored outcomesReady 5.3 High-dimensional omics and feature selectionReady 5.4 Missing data, imbalance and fairnessReady 5.5 Final applied ML project in RReady

Open module →

Recommended path

Start with Module 1 before moving to supervised learning.

Module 1 gives students the essential judgement needed for the rest of the course: prediction thinking, causality caution, learning types, train/test validation, leakage prevention and workflow-based reporting.

Open Module 1 →Preview Module 2 →