Hyperparameter selection

Secondary Analysis of Electronic Health Records (pp 419-427)

Published September 10, 2016

Franck Dernoncourt, Shamim Nemati, Elias Baedorf Kassis, Mohammad Mahdi Ghassemi

Algorithms and features in medical studies contain many “knobs” that govern the learning process from a high-level perspective: they are called hyperparameters, and investigators typically tune them by hand. In this case study, we present three mathematically grounded techniques to automatically optimize hyperparameters, and demonstrate their use in the problem of outcome prediction for ICU patients who suffer from sepsis.

Learn More

Research Area:  AI & Machine Learning