Machine Learning for Health Applications, Zimbabwe 2026

Welcome to the Machine Learning for Health module.

This course introduces the foundations of machine learning, with a strong focus on:

• Linear models
• Regularisation
• Model evaluation
• Ensemble methods
• Application to real-world health data

The course builds from first principles and culminates in applying machine learning methods to health-related prediction tasks.

---

Teaching Team

Iona Biggart AI 4 Paediatrcis PhD student at Imperial College London (iona.biggart23@imperial.ac.uk)

Marco Reed AI 4 Paediatrcis PhD student at Imperial College London (marco.reed24@imperial.ac.uk)

Kevin Meck Biomedical and applied medical AI engineer at Neotree

Aditi Rao Doctoral candidate at Imperial College London with Neotree

Guest Lectures

Professor Payam Barnaghi Chair in Machine Intelligence Applied to Medicine, Imperial College London

Professor Tawanda Mushiri Associate Professor, AI and Robotics, SIRDC

Antigone Fogel PhD student, AI for dementia research, UK DRI and Imperial College London

---

Course Overview

Machine learning is transforming modern healthcare.
From disease prediction to risk stratification and treatment modelling, robust statistical learning methods are central to evidence-based medicine.

This course focuses on core machine learning methods that form the backbone of applied health data science.

We concentrate on:

1️⃣ Linear Models

• Linear regression
• Logistic regression
• Bias–variance tradeoff
• Regularisation (L1 / L2)
• Model interpretation
• Feature importance
• Evaluation metrics

Linear models are essential because:

• They are interpretable
• They are statistically grounded
• They often perform surprisingly well in health data

---

2️⃣ Ensemble Methods

• Decision trees
• Random forests
• Gradient boosting
• Model comparison

Ensemble models:

• Capture non-linear structure
• Improve predictive performance
• Handle complex feature interactions

---

3️⃣ Application to Health Data

The final part of the course applies these methods to:

• Clinical-style datasets
• Structured tabular health data
• Real-world prediction problems

You will:

• Preprocess datasets
• Train models
• Compare performance
• Justify modelling decisions
• Interpret results in a healthcare context

---

Learning Objectives

By the end of this course, you should be able to:

• Understand the mathematical foundations of linear models
• Explain regularisation and its necessity
• Implement regression and classification models in Python
• Compare model performance using appropriate metrics
• Apply ensemble models to structured datasets
• Critically evaluate modelling choices in health applications

Guest lectures:

• Kevin Meck: Neotree, sepsis modelling
• Aditi Rao: Neotree, implementation of ML tools in clinical settings
• Antigone Fogel: Predictive modelling for dementia progression

---

Repository Structure

This repository is organised into clearly structured folders:

📂 lectures/

Contains:

• Lecture slides
• Supplementary notes

📂 labs/

Contains:

• Jupyter notebook tutorials
• Practical exercises
• Applied modelling workflows

Each lab can be run locally or opened directly in Google Colab.

📂 assets/

Contains images and supporting files for documentation.

---

Running the Labs

You may run the notebooks:

Option 1: Locally

Create a Python environment:

conda create -n mlhealth python=3.11
conda activate mlhealth
pip install -r requirements.txt

Afternoon Labs

Day 1 – Python for Beginners

Click below to open in Google Colab:

Day 2 – SVM, decision trees and random forest

Day 3 - additional labs

Linear algebra:

Normalisation and standardisation:

Machine learning for beginners assessment (optional):

Course Origins

This module has been adapted and streamlined from:

Machine Learning for Neuroscience (ML4NS)
https://ml4ns.github.io/

The original ML4NS course provided a comprehensive introduction to machine learning with applications in neuroscience.

This version has been redesigned to:

• Focus more heavily on linear and ensemble methods
• Streamline deep learning components
• Emphasise health and clinical data applications

We gratefully acknowledge the original ML4NS teaching materials and contributors.

---

Acknowledgements

We thank all past contributors to the ML4NS course and the Translational Machine Intelligence Lab for foundational teaching materials. We thank Dr. Felicity Fitzgerald and Professor Michelle Heys for their support and collaboration.

Funding

This course was made possible by support from the organisations listed above as well as the Imperial College London Dean's PhD Professional Development Award granted to Iona Biggart and Marco Reed.