Method for Dynamic Simulation Parameter Calibration by Machine Learning

摘要

The present invention relates to a method for calibrating a dynamic simulation parameter based on machine learning performed by a computing device. The method comprises the steps of: generating a set of the N number of parameter hypotheses; obtaining result values corresponding to each of the N number of parameter hypotheses; calculating likelihoods corresponding to each of the result values; applying a Hierarchical Dirichlet Process Hidden Semi-Markov Model (HDP-HSMM); obtaining regime search result values; obtaining maximum likelihood estimation data for each regime by applying a maximum likelihood estimation method based on the regime search result values; and determining a maximum likelihood parameter based on the maximum likelihood estimation data.