基于EEMD-SSA-LSSVR的短期交通流预测  

Short Term Traffic Flow Prediction Based on EEMD-SSA-LSSVR

作　　者：李俊 胡婷 LI Jun;HU Ting(School of Business Administration,Chongqing Technology and Business University,Chongqing 400067,China)

机构地区：[1]重庆工商大学工商管理学院,重庆400067

出　　处：《现代信息科技》2022年第6期93-96,共4页Modern Information Technology

基　　金：重庆工商大学研究生创新型科研项目(yjscxx2021-112-14、yjscxx2021-112-15)。

摘　　要：为了提高物流的配送效率,寻求合理的配送路径,通过分析短期历史交通流量,使用集合经验模态分解去噪,以拟合优度最大化为目标,运用麻雀搜索算法优化惩罚参数和核函数参数的最小二乘支持向量机回归预测短期交通流。结果表明集合经验模态分解能有效去除短期交通流中的噪声,构建的EEMD-SSA-LSSVR模型可以高效地预测短期交通流量。To improve the distribution efficiency of logistics and find the reasonable distribution route,the ensemble empirical mode is used to decompose noise signals by analyzing the short-term historical traffic flow.Aiming at maximizing the goodness of fit,the sparrow search algorithm is used to optimize the least squares support vector machine regression of penalty parameters and kernel function parameters to predict the short-term traffic flow.The results show that the ensemble empirical mode decomposition can effectively remove the noise in the short-term traffic flow,and the constructed EEMD-SSA-LSSVR model can effectively predict the short-term traffic flow.

关 键 词：麻雀搜索算法 集合经验模态分解 短时交通流预测 最小二乘支持向量机回归 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]