Application of Elman neural network based on improved niche adaptive genetic algorithm

The structure and algorithm of an ordinary improved Elman neural network is introduced at first. In order to overcome the problem of being trapped into local minimal points and slow convergence, an improved niche adaptive genetic algorithm was proposed in this paper, which can simultaneously optimize the weights and self-feedback gain coefficient, finally it was applied to the modeling of the near infrared spectrum of cashmere and wool. The near infrared spectrum of cashmere and wool is a nonlinear, strong coupled system, and traditional methods can hardly build up its mathematics model. The simulation experiment based on the cashmere and wool from various regions indicate that the Elman neural network based on improved niche adaptive genetic algorithm make a good match for the near infrared spectrum of cashmere and wool, and it has better dynamic performance, quicker approach speed, better precision and generalization ability.