Machine learning models for CF

Project 1: Bayesian inference with Expectation Maximisation for the characterisation of antibiotic treatment recovery in Cystic Fibrosis

Project 2: Estimation of the variability in Cystic Fibrosis patient’s FEV1 lung function measurements

Hosted at the University of Cambridge

Documentation

The code documentation is available here.

Repository structure

    ├── FEV1variability               <- code base for the side project
    ├── docs                          <- code documentation
    ├── exploration                   <- code base for exploration of the data
    ├── helperfunctions               <- helper functions common to at least two other matlab scripts
    ├── msc-tristan                   <- important documents necessary to reproduce all results
        ├── report                      <- thesis report with LaTeX version, thesis defense presentation
    ├── recovery                      <- code base for the main project
        ├── updatedModel                <- updated version of the ML model

Code base usage

1. Clone this repo in a folder called Code/
2. In that same folder, clone this version of the repo - it shows the project version at commit 9179127f15e96085db14e362f9d43b36f488472e (26.07.2021)
3. You should obtain the following folder structure Code/master_thesis_CF_ML and Code/smartcare
4. Change the absolute pathes in the init.m files - see FEV1variability, exploration, recovery
5. You are ready!

Main project

1. Run masterscriptRecovery to load the data
2. Explore data with functions in exploration
3. For the main project: run an alignment model with runAlignmentModelEMMCRecoveryFcn

Side project

1. Run analyseFEV1Variability

Data

The data is not available as it is stored in a secured server due to privacy reasons.

Contact

Contact me at [email protected] if you would like to know more about the project!