BACKGROUND
The current investigations are related to design a novel stochastic solver based on the radial basis Bayesian regularization neural network (RBBRNN) for the fractional bone transformation model (BTM).
OBJECTIVE
The fractional derivatives have been presented to get more precise and accurate performances of the BTM using the stochastic computing performances of the RBBRNN. The exactness of the novel designed RBBRNN is perceived through the comparison of the achieved and reference solutions. The fractional BTM is categorized into three classes, osteoclasts, osteoblasts, and tumor cell densities.
METHODS
Three different cases based on the fractional BTM are numerically presented by using the RBBRNN through the training and testing procedures to decrease the mean square error.
RESULTS
Thirteen neurons in the hidden layers for solving the fraction BTM are presented along with input/output layers performances.
CONCLUSIONS
The competence of the designed RBBRNN is presented using the correlation/regression, state transitions and error histograms.