Bilinear Neural Network Construction for a Fractional Konopelchenko-Dubrovsky-Kaup-Kupershmidt Equation

This paper solves a fractional generalized (3+1)-dimensional Konopelchenko-Dubrovsky-Kaup-Kupershmidt equation using the bilinear neural network method to construct Hirota bilinear expression. The technique demonstrates solutions in terms of sigmoid activation functions, softmax activation functions, and periodic solutions. The graphs of the solutions are in 3D and one-dimensional contour to investigate the behavior. Compared to the previous research, the BNNM is a practical, adequate, and convenient technique for obtaining solutions by constructing the activation functions through a recurrent neural network model.