Figure 1
From: Prognostication and Risk Factors for Cystic Fibrosis via Automated Machine Learning
Schematic depiction of the AutoPrognosis framework. AutoPrognosis is provided with a dataset and a definition for an appropriate clinical utility selected by clinical experts. The algorithm uses Bayesian optimization in order to update its beliefs about the clinical utility of different machine learning pipelines, where each pipeline comprises an imputation algorithm, a feature processing algorithm, a classification algorithm and a calibration method. In this depiction, a pipeline comprising MICE imputation, fast ICA processing, XGBoost classifier and sigmoid calibration is highlighted.
