A new algorithm for calibrating agent-based models is proposed, which employs a popular gradient boosting framework. Machine learning techniques are not used to develop a surrogate model, but rather assist in narrowing down the parameter space during the search for optimal parameters. Our...
