Numerical investigation of a human liver model based on generalized fractional operators and neural networks

In this work, we analyze human liver function using a modified Atangana-Baleanu-Caputo (MABC) fractional derivative with respect to another function in a mathematical model (MMD) to better understand and predict liver behavior. We employ iteration methods and fixed-point theory to establish the existence of a unique solution for this model. Additionally, we apply the homotopy transform method to solve the MMD and evaluate its accuracy. Numerical experiments validate the results, with this research aiming to enhance the precision of liver damage assessment and early detection, ultimately contributing to improved treatment strategies. Graphical representations and tabular data illustrate comparisons between our proposed model and existing approaches. The mean square error (MSE), gradient error histogram, and actual versus predicted values of the solution are examined using an Artificial Neural Network (ANN). The new fractional model outperforms traditional integer-order models that use standard temporal derivatives based on comparisons with actual clinical data.