The Analysis And Predict Model For Traffic Flow Magnitude Basised On SVM

The support vector machine is a new kind of machine learning algorithm, It takes the statistical study theory as its base theory. Its training equally in value to solves one two plans problems, using the structure risk smallest principle, has the forecast ability is strong, the overall situation is most superior and the convergence rate is quick and so on the remarkable characteristic. Therefore, this article obtains from the research chaos time series forecast, proposed one kind based on the support vector machine return theory forecast method, and applied it in the chaos time series forecast. In order to confirm this algorithm the performance, we have done two work, one was the use space construction theory carries on to the chaos time series on foot in the many steps of forecasts, and carried on the comparison analysis with the literature in nerve network forecast result; the other is to join the different horizontal noise in the chaos time series, and carried on the comparison analysis with the BP network and the RBF network forecast result. The simulation result indicates that carries on the chaos time series with the support vector return algorithm the forecast can obtain better effects compared to other method even, has the very good steadiness and exudes the ability.With the foundation in chaos time series forecast research, we make the study about highway traffic flow magnitude forecasts according to the characteristic of highway traffic flow magnitude, and we established the highway traffic flow magnitude short-term forecast model based on the support vector machine, applied the actual load data carries on the simulation experiment, study indicated that this method has obvious advantage over the artificial nerve network method in the forecast effect.