Research on Improved BP Neural Network Prediction Based on Niche Particle Swarm Optimization Algorithm

Abstract

As for data prediction, a prediction model is proposed based on BP neural network. This method is based on the traditional BP neural network algorithm to improve the problems of slow convergence speed and easy to fall into local optimum. The idea of niche is introduced into the particle swarm optimization algorithm to improve the global search ability and convergence speed, find the global optimum value, and then use LMBP neural network to continue to optimize the parameters to improve the prediction effect.