Evolutionary-based Hyperparameter Tuning in Machine Learning Models for Condition Monitoring in Wind Turbines – A Survey

摘要

Optimality of model hyperparameters is essential for intelligent condition monitoring (ICM) of wind turbines using machine learning models, hence the need for hyperparameter tuning. Evolutionary algorithms (EAs) have been used for hyperparameter tuning of machine learning models, however, little is known about the hyperparameter tuning of these EAs. This study presents a survey of hyperparameter tuning of EAs used for tuning hyperparameters of machine learning models that are used in ICM of wind turbines. Findings show that many studies tune hyperparameters for machine learning models, however, a few studies tune these hyperparameters with EAs. Among these few, a handful tune the hyperparameters of such EAs and such studies in ICM of wind turbines is very sparse. Hence the need to explore this double stage hyperparameter (DSHP) tuning in ICM of wind turbines.