Advanced Machine Learning Algorithms for Accurate Prediction of Band Gaps in Rare Earth Metal Oxides Nanoparticles
Abstract
The accurate prediction of band gaps in rare earth metal oxides nanoparticles is crucial for their application in various fields, including catalysis, solar cells, electronics, waste water treatments, and solid oxides fuel cells. This study presents reports the machine learning algorithms to predict the band gaps of cerium dioxide (CeO2) and neodymium oxide (Nd2O3) nanoparticles with high precision. Band gaps data were obtained from experimental and computational sources using UV-visible spectrophotometry and applying Wood and Tauc equation. Seven Machine Learning (ML) models including Random Tree (RT), Linear Regression (LR), Instance-Based KNN (IBK-KNN), Multilayer Perceptron (MLP), Random Forest (RF), Support Vector Machine –Regression (SVR) and Additive Regression (AD) were employed. The models were trained using these ML models and validated using rigorous cross-validation techniques to ensure their generalizability and reliability. The findings demonstrate that the machine learning models significantly performed by achieving the correlation coefficient of 1.00, a mean absolute error (MAE) below 0.01 eV. The developed ML model not only enhances the understanding of band gap behavior in CeO2 and Nd2O3 but also provides an optimization of band gap values of rare earth metal oxides nanomaterials.