Research On BP Neural Network Algorithms Based On Intelligent Algorithms

BP neural network is widely used in numerical prediction and pattern recognition because of its strong nonlinear mapping ability,self-learning and self-adaptive ability.But the traditional BP algorithm is easy to fall into local optimum because of random selection of initial weights and thresholds,while the intelligent algorithm has strong optimization ability,therefore,it is of extremely great practical significance to use intelligent algorithm to study BP neural network algorithm.This paper introduces the research background,significance,research status and principles of BP neural network,and focuses on the improvement of BP neural network algorithm by intelligent algorithm.Firstly,a BP neural network algorithm based on cuckoo algorithm is set up.Aimed at the shortcomings of the original cuckoo algorithm,a cuckoo algorithm based on function dynamic decline factor(ICS)is proposed,Three typical test functions are used to verify the effectiveness of the improved algorithm.Then,the algorithm is used to find the initial weights and thresholds of BP network,and it is applied to predict the price of commodity houses in China.The simulation results show that the prediction accuracy of ICS-BP network model is better than that of CS-BP network model and logistic prediction model.Secondly,a BP neural network algorithm based on bat algorithm is set up.Aimed at original bat algorithm shortcomings of slow convergence speed and easy to fall into local minimum,an adaptive speed and mutation bat algorithm(ADBA)is proposed,six typical test functions are selected to simulate ADBA and BA respectively.The simulation results show that ADBA is superior to BA in both precision and convergence speed.Then the algorithm is used to optimize the initial weights and thresholds of BP network,and it is applied to population prediction in China.The results show that ADBA-BP network model is superior to BA-BP network model and logistic prediction model in prediction accuracy.Finally,the content of this article is summarized and further work is expected in the future.