Comparative Analysis of Support Vector Machine Regression and Gaussian Process Regression in Modeling Hydrogen Production from Waste Effluent

Organic-rich substrates from organic waste effluents are ideal sources for hydrogen production based on the circular economy concept. In this study, a data-driven approach was employed in modeling palm oil mill and activated sludge waste. Seven models built support vector machine (SVM) Gaussian process regression (GPR) were of sources. The SVM incorporated with linear kernel function (LSVM), quadratic (QSVM), cubic (CSVM), fine (GFSVM). While GPR rotational (RQGPR), squared exponential (SEGPR), (EGPR). model performance revealed that SVM-based did not show impressive effluent, as indicated by R2 ?0.01, 0.150, 0.143 LSVM, QSVM, CSVM, respectively. Similarly, perform well sludge, evidenced values 0.040, 0.190, 0.340 On contrary, SEGPR, RQGPR, EGPR displayed outstanding prediction both effluent over 90% datasets explaining variation output. With > 0.9, predicted consistent minimized errors. level importance analysis all input parameters relevant hydrogen. However, influent chemical oxygen demand (COD) concentration medium temperature significantly influenced whereas pH sludge.