关键词: 搜索者优化算法, 最小二乘支持向量机, 短时交通流量, 预测

Abstract: In view of the nonlinear and uncertainty for shortterm traffic flow characteristic, a forecasting model based on SOA-LSSVM is proposed, combining SOA’s advantages of high convergence precision with LSSVM’s superiority in fast solving speed. SOA is used in choosing regular parameter and nucleus parameter for LSSVM, thus, a optimal forecast model is established based on SOA-LSSVM. The model was used in shortterm traffic flow prediction, and the application results shows that compared with traditional ANN, SOA-LSSVM model has a better prediction precision and application effect, and this model is fit for shortterm traffic prediction with a good popularization value.

Key words: seeker optimization algorithm, least squares support vector machines, shortterm traffic flow, prediction