果蝇优化最小二乘支持向量机混合预测模型——以我国物流需求量预测为例

A Hybrid Forecasting Model Based on Fruit Fly Optimization Algorithm and Least Squares Support Vector Machine:——the Case of Logistics Demand Forecasting of China

为解决最小二乘支持向量机参数设置的盲目性,利用果蝇优化算法对其参数进行优化选择,进而构建了果蝇优化最小二乘支持向量机混合预测模型.以我国物流需求量预测为例,验证了该模型的可行性和有效性.实例验证结果表明:与单一最小二乘支持向量机和模拟退火算法优化最小二乘支持向量机预测模型相比,该模型不仅能够有效选择参数值,而且预测精度更高.

To solve the blindness of parameter settings of least squares support vector machine （LSSVM）, the fruit fly optimization algorithm （FOA） was used for the optimal selection of the parameters, then a hybrid forecasting model （FOALSS- VM） based on FOA and LSSVM was proposed. Taking the logistics demand forecasting of China as example, the feasibility and effectiveness of the FOALSSVM model were proved. Case study shows that the FOALSSVM model, compared with the single LSSVM model and LSSVM model optimized by simulated annealing algorithm （SALSSVM , can effectively choose the parameters values and improve the forecasting accuracy.