Logistics demand forecasting based on LSSVM optimized bytwo-order oscillating PSO
基于二阶振荡微粒群最小二乘支持向量机的物流需求预测

Based on analyzing the factors of logistics demand, this paper proposed a new model named the two-order oscillating particle swarm least squares support vector machines TOOPSO-LSSVM model to improve the forecasting accuracy of logistics demand. The complex nonlinear relationship between logistics demand and its impact factors were explained through LSSVM. And then, it used TOOPSO algorithm to optimize the parameters of LSSVM model. An empirical analysis indicates that the forecasting performance of LSSVM is better than the other three models and the searching time for optimal parameters of LSSVM by TOOPSO is obviously less than cross validation method, which is an effective method for logistics demand forecasting.