The research work will focus on the evaluation of the ecological pressure with two features, namely ELR-Class for categorical forms of ecological state and ELR for continuous forms of ecological loads. The research work will employ supervised classification and Gaussian-based clustering on ELR-Class, while ELR will employ supervised regression models for forecasting. The experiments will show the accuracy of supervised classifiers compared to the accuracy of the unsupervised clustering model, registering more than 0.98 accuracy for Gradient Boost, XGBoost, and AdaBoost classifiers on ELR-Class. The accuracy of ELR forecasting models, namely ensemble regressors ABR, LGBR, and XGBR, registered R² values above 0.97.
Multi-Model Machine Learning Analysis of Ecological Load Ratio Using Classification, Regression, and Clustering Approaches in EU Countries
12 Views
3 Downloads