Support Vector Regression And Its Application Based On Intelligent Optimization Algorithm

The paper analyzes the performance of support vector regression (SVR) with various parameters which also have different influences and the shortcomings. In order to further improve the forecasting accuracy, aiming at the deficiency in the parameters selection of SVR, the selection of parameters is considered as a compound optimization problem in view of the integrative influence by each parameter and the objective function is set. Then, three intelligence optimization techniques (Genetic Algorithm; Particle swarm optimization; Ant colony optimization) have been employed into the optimization of parameters for forecasting model respectively based on SVR that makes automatically selection of parameters come true.In practice, GA-SVR model is drawn on the coal demand forecasting, PSO-SVR model is utilized to the foodstuff output forecasting, ACO-SVR model is applied to the regional logistics volume forecasting, by means of MATLAB and the toolbox of LIBSVM to writing the corresponding program, obtain the results and draw the curve of results. The simulation results demonstrate that compared with conventional method, the proposed methods can reduce modeling error and forecasting blunder of SVR model effectively and have superior learning and generalization performance. Besides, they are convenient and fast for calculating and programming. Thus, the approaches have higher practicability as well as play three intelligent optimization algorithm optimization capabilities in their own characteristics and merits.