Application of Improved Support Vector Machine Algorithm in Short-term Traffic Flow Forecast

The paper introduces cross-validation and grid-search method to optimize the prediction accuracy of Support Vector Machine models,the establishment of an improved Support Vector Machine prediction model,and applied to short-term traffic flow forecasting empirical analysis.The paper also uses the real time data of certain urban road to test the efficiency of the proposed model and the result is satisfactory.