On the investigation of the dynamical mechanisms of neuroblastoma by stochastic neural networks

To comprehend the interactions between cancer cells, tumor-suppressor cells, immune-suppressor cells, and oncolytic viruses, the dynamical functions of neuroblastoma are described using the mathematical neuroblastoma model (NBLM). Further, mathematical study is computationally carried out by using the Bayesian-Regularization Algorithm (BRA) based on Stochastic Neural Networks (SNNs), i.e., BRA-SNNs to verify the realistic effects of neuroblastoma disease. The BRA-SNNs are implemented on datasets that are generated by using the Runge Kutta Method (RKM) and segmented into testing, training, and validation sets to get approximate outcomes. Furthermore, the efficiency, precision, and accuracy of BRA-SNNs are analyzed through Absolute Errors (AEs) and Mean-Square Errors (MSEs). The performance, efficiency, convergence, and reliability of the implemented algorithm are checked depending on the attained outcomes of training-state analysis, regression analysis, and error-histogram analysis.