Annual Runoff Forecast Based on Cooperative Particle Swarm Projection Pursuit Regression Model

According to the high-dimensional nonlinear problem of annual runoff prediction, to build runoff forecasting model based on projection pursuit regression model of Hermite polynomials and the cooperative particle swarm optimization algorithm. Projection pursuit prediction model projects high-dimensional data into low-dimensional space based on sample data driving, completely according to the sample data driven to enhance the prediction results objectivity. The particle swarm optimization algorithm combines the idea of co-evolution to optimize the projection direction and polynomial coefficients in parallel, and further improve the convergence rate and prediction accuracy of the model. The model is applied to the flow prediction of Jiubujiang River Reservoir. The relative error of runoff prediction is less than 15%, and the prediction result is high precision and reliability. The experimental results show that it is feasible and effective to use the cooperative particle swarm projection pursuit regression model to predict the annual runoff.