The PREdiction by SUpervised Learning Toolkit (PRESULT) is designed to simplify the development, validation and optimization of machine learning (ML) risk prediction models. PRESULT supports four machine learning methods: Random Forest, Mixture of Experts, Support Vector Machine, and Regression Tree. The tool produces portable models and program objects that can be distributed and used for validation or prediction on new datasets. PRESULT also generates receiver operating characteristics (ROC) curves and calculate the area under the curve (AUC) for testing, training and k-fold cross-validation datasets. Finally, PRESULT can apply trained machine learning models to predict individualized absolute risk of developing disease over one or more years.

Executables for Mac, Linux, and Windows are available for download, as well as the user guide and source code. More information will be available upon manuscript publication.