Optimization of Grey Neural Network Model Based on Mind Evolutionary Algorithm

For the problem of randomized parameters and large error, the grey neural network model optimized by mind evolutionary algorithm(MEA-GNNM) was presented. The experiment results showed that the total error, root mean square error, mean percentage error, and running time of MEA-GNNM were 4457.6000, 920.7372, 4.04% and 1.1509 seconds; Compared with GA-GNNM, the total error, root mean square error and mean percentage error of MEA-GNNM had decreased by 16.12%, 12.39% and 16.87%, and the calculation speed increased by 21.61%. Compared with PSO-GNNM, the total error, root mean square error and mean percentage error of MEA-GNNM had decreased by 9.13%, 11.33% and 12.17%, but running time increased by 104.50%. The paper presents a new approach to optimize the parameters of grey neural network model, and also provides a new method having higher prediction accuracy for the time series prediction.