Research On Improved PSO-BP Combination Forecasting Model And Application

Data prediction is a hot topic in academic research,mainly through the analysis of existing data,to achieve the purpose of forecasting future data.When the data is predicted,it is necessary to achieve the corresponding prediction effect,so that it has practical significance.Predictive effects depend on the predictive model,and the predictive model is studied to improve the accuracy of the prediction,which is very meaningful.The existing prediction model can be divided into single prediction model and combined forecasting model.The BP neural network in the single prediction model can simulate the nonlinear relationship between the data and the prediction effect is better,but the BP neural network itself has some defects.In this paper,the BP neural network optimization method is analyzed and studied.On this basis,the process of PSO optimization BP is improved,and an improved PSO-BP model is proposed.The improved model uses the BP neural network to update the optimal position of the population and the optimal position of the particles in the PSO optimization BP process.The experimental results show that the improved PSO-BP model is faster and the prediction effect is better than that of PSO-BP.The combined forecasting model can absorb the advantages of each single prediction model,and there is no case where the prediction effect is very poor at some point.Taking into account this,this paper uses the Adaboost algorithm to combine the improved PSO-BP,SVM and gray prediction models to propose a new combination forecasting model.This will not only improve the advantages of improved PSO-BP,but also with other predictive models to complement each other,resulting in more accurate predictions.The new combination forecasting model is used in the R & D input prediction experiment.Compared with the single prediction model and other combination forecasting model,the results show that the new combination forecasting model has the best prediction effect and the prediction error is the smallest.In this paper,an improved PSO-BP model is proposed and combined with SVM and gray prediction model.A new combination forecasting model is proposed.The new combination forecasting model is used to predict the experiment.The experimental results show that the combined forecasting model is more effective and accurate high.