Program processors;Feature extraction;Measurement;Meteorology;Anomaly detection;Event detection;Decision trees;I.6.6 [Computing Methodologies]: Simulation and Modeling-Simulation Output Analysis;I.5.0 [Computing Methodologies]: Pattern Recognition-General

Using feature importance metrics to detect events of interest in scientific computing applications

With current high performance scientific computing workflows, data are typically recorded at regular intervals spaced several hundred time steps apart. Data are not saved at every time step to prevent excessive memory usage and because data I/O is …