摘要
针对当前智慧城市交通拥堵治理中的车流量预测准确度低、预测效果差的问题,提出了基于LSTM-LSSVM的车流量预测方案,在历史车流量样本数据中,通过LSTM和LSSVM对车流量划分成为的训练集使用不同的拟合方式改进了预测效果。实验中,以某一个区域的历史车流量为研究样本,预测指标结果说明该方法相比LSTM、LSTM-SVM具有较好的预测效果。
For the current problems of low accuracy and poor effect of traffic flow prediction in traffic congestion management in smart cities,a traffic flow prediction scheme based on LSTM-LSSVM is proposed.And in the historical traffic flow sample data,it uses different fitting methods to improve the prediction effect for the training set which is divided by LSTM and LSSVM according to the traffic flow.In the experiments,the historical traffic flow of a region is used as the research sample,and the results of the prediction indexes show that the method has better prediction effect compared with LSTM,LSTM-SVM.
作者
陈暄
CHEN Xuan(Zhejiang Industry Polytechnic College,Shaoxing 312000,China)
出处
《现代信息科技》
2022年第21期109-111,共3页
Modern Information Technology
基金
绍兴市哲学社会科学研究“十四五”规划2022年度重点课题—指南课题(145192)。
关键词
智慧城市
车流量
预测
smart city
traffic flow
prediction
