ABSTRACT This study investigates the influence of nonlinear thermal and solutal buoyancy forces, together with inclined magnetic fields, on mixed convection boundary‐layer flow of a viscoelastic fluid over a vertically stretching permeable plate, while also assessing the predictive capability of artificial neural networks (ANNs) for heat transfer and flow characteristics under nonlinear effects. A collocation‐based solver (MATLAB bvp5c) is employed to numerically analyze the governing boundary layer equations under both linear and quadratic Boussinesq approximations, with three radiation models (linear, quadratic, and nonlinear). Parametric studies are conducted to evaluate the roles of the magnetic field inclination angle, viscoelasticity and radiation parameter on velocity, temperature, concentration, Nusselt number, and skin friction, while an ANN model is trained on simulation datasets using different training algorithms and evaluated through a newly proposed composite index combining , RMSE, and sMAPE. The results reveal that nonlinear thermal and solutal buoyancy effects substantially modify the velocity, temperature, and concentration fields. The ANN model demonstrates excellent predictive performance along with sensitivity analysis, with Bayesian regularization emerging as the most reliable training algorithm according to the composite evaluation metric. Overall, sensitivity analysis of the Nusselt number indicates that the radiation parameter has the strongest influence on heat transfer, whereas the viscoelastic parameter exhibits the least effect.
Rawal et al. (Tue,) studied this question.