TY - JOUR
T1 - Artificial intelligent investigations for the dynamics of the bone transformation mathematical model
AU - Cholamjiak, Watcharaporn
AU - Sabir, Zulqurnain
AU - Raja, Muhammad Asif Zahoor
AU - Sánchez-Chero, Manuel
AU - Gago, Dulio Oseda
AU - Sánchez-Chero, José Antonio
AU - Seminario-Morales, María Verónica
AU - Gago, Marco Antonio Oseda
AU - Cherre, Cesar Augusto Agurto
AU - Altamirano, Gilder Cieza
AU - Ali, Mohamed R.
N1 - Publisher Copyright:
© 2022
PY - 2022/1
Y1 - 2022/1
N2 - In this study, the stochastic numerical solutions of the fractional myeloma bone disease system (FMBDS) have been presented. The fractional order investigation provides more accurate solutions of the FMBDS. The FMBDS is classified into three dynamics and the solution of each class is presented by using the artificial neural network enhanced by the scale conjugate gradient procedures (ANN-SCGPs). Three different fractional order performances have been used to present the solutions of the FMBDS by applying the ANN-SCGPs. The statics is chosen as 11%, 12% and 77% for training, testing and verification. Twelve number of hidden neurons with input and output layers have been proposed for the FMBDS. The comparison of proposed and reference solutions is performed that shows the accuracy of the ANN-SCGPs. The consistency, validity, precision, and capability of the ANN-SCGPs can be judged based on the state transitions values, regression actions, correlation behaviors, error histograms, and mean square error data.
AB - In this study, the stochastic numerical solutions of the fractional myeloma bone disease system (FMBDS) have been presented. The fractional order investigation provides more accurate solutions of the FMBDS. The FMBDS is classified into three dynamics and the solution of each class is presented by using the artificial neural network enhanced by the scale conjugate gradient procedures (ANN-SCGPs). Three different fractional order performances have been used to present the solutions of the FMBDS by applying the ANN-SCGPs. The statics is chosen as 11%, 12% and 77% for training, testing and verification. Twelve number of hidden neurons with input and output layers have been proposed for the FMBDS. The comparison of proposed and reference solutions is performed that shows the accuracy of the ANN-SCGPs. The consistency, validity, precision, and capability of the ANN-SCGPs can be judged based on the state transitions values, regression actions, correlation behaviors, error histograms, and mean square error data.
KW - Artificial neural networks
KW - Bone transformation
KW - Cell dynamics
KW - Fractional order
KW - Myeloma disease
KW - Scale conjugate gradient procedures
U2 - 10.1016/j.imu.2022.101105
DO - 10.1016/j.imu.2022.101105
M3 - Artículo
AN - SCOPUS:85139723062
SN - 2352-9148
VL - 34
JO - Informatics in Medicine Unlocked
JF - Informatics in Medicine Unlocked
M1 - 101105
ER -