Upon publication, we will publish the code for our article titled "Deep learning-based patient stratification for prognostic enrichment of clinical dementia trials" in this repository.
All outputs and pointers towards the patient data have been removed for patient data privacy. Therefore, the code is not executable in its current form.
The code for training a deep learning model for clustering the patient trajectories can be found here.
- src/: Hold all the source code
- clustering/: Code for clustering, cluster validation, and replication.
- classifier_training/: Nested cross-validation, Bayesian hyperparameter optimization, xgboost classification model
- trial_comparison/: Code for the power analysis and the cost analysis presented in the article.