Regression analysis for thermal transport of fractional-order magnetohydrodynamic Maxwell fluid flow under the influence of chemical reaction using integrated machine learning approach

Abstract

An innovative idea of regression analysis based on machine learning technique for magnetohydrodynamic flow of Maxwell fluid within a cylinder is proposed. Mean Squared Error is used for the simulation of heat transfer and fluid flow. The governing flow equations involving a system of coupled, nonlinear fractional partial differential equations are solved by homotopic approach called HPM. The predicted solution is obtained with Python built-in code on Google-Colab. The effects of Atangana-Baleanu fractional time order derivative on the momentum, thermal, and concentration boundary layer are analyzed. It is observed that the momentum boundary layer gets higher and higher by increasing the values of Atangana-Baleanu fractional time order derivative. The thermal boundary layer shows improvement with the increasing value of the Peclet number. The concentration boundary layer thickness declines with the growing values of chemical reactions. The validation of results is examined by MSE, function fit, and correlation index.