Performance Prediction of Carbon-Based Supercapacitor Electrodes as Energy Storage Devices Using Machine Learning Techniques

Abstract

Ongoing years have seen the expansive utilization of carbon electrodes for electric double layer capacitors (EDLCs) due to enormous surface region, high porosity and minimal expense. Carbon is the most broadly involved electrode for the supercapacitors. This work applies the different machine learning technologies to predict the capacitance of carbon-based supercapacitors. As the result, comparing to other machine learning methods, such as ridge regression (RR), K-nearest-neighbor (KNN), decision trees (DTs), Bayesian linear regression (BR) and support vector machines (SVMs), KNN exhibits the best accuracy and adaptability in the capacitance predication and i.e., KNN provides a root mean square error (RMSE) of approximately 83.945.