Hyperparameter optimization (HPO) in machine learning pipelines can significantly impact model performance. While traditional approaches such as manual tuning, grid search, and random search are widely adopted, they become inefficient when dealing with complex, high-dimensional search spaces. To address this limitation, this paper introduces mloptimizer , an open-source package that implements genetic algorithm-based HPO for machine learning models. The package integrates seamlessly with scikit-learn , the widely used Python machine learning framework, and offers compatibility with popular libraries such as XGBoost and CatBoost . Built on top of the DEAP (Distributed Evolutionary Algorithms in Python) library, mloptimizer supports parallel execution for efficient exploration of large search spaces. In addition, the package is designed to be easy to use and highly customizable, and it provides native integration with MLflow for experiment tracking and reproducibility.
Caparrini et al. (Fri,) studied this question.