Identify predictive parameters for 28-day mortality using a genetic algorithm (GA) and Gaussian mixture models (GMM).
##Design Identify predictive variables for patient cohort using a unsupervised data drive process.
##Cohort All adult ICU admissions using MIMIC (http://mimic.physionet.org)
##Methods/Software SQL (Oracle) to extract data R to run Genetic Algorithm (GA) and Expectation-Maximization (EM) for Gaussian Mixture Model (GMM)