基于CNN-BiLSTM的短时公交客流量预测  

Short term bus passenger flow prediction based on CNN-BiLSTM

作　　者：石秀君[1] 李鹏浩[1] SHI Xiujun;LI Penghao(Jining Polytechnic,Jining 272000,China)

机构地区：[1]济宁职业技术学院,山东济宁272000

出　　处：《中国高新科技》2023年第22期129-130,共2页

摘　　要：文章以实现公交客流量预测为目标,利用深度学习技术建立基于卷积神经网络与双向长短期神经网络的公交客流量预测模型(CNN-BiLSTM),根据历史交通数据和当前状态进行预测公交客流量,并与其他客流量预测模型的结果进行对比。结果表明,CNN-BiLSTM客流量预测算法的RMSE和MAE最小,预测精度相对较高。This article uses deep learning technology to establish a public transportation passenger flow prediction model(CNN BiLSTM)based on a combination of convolutional neural networks and bidirectional short-term and short-term neural networks.The model predicts public transportation passenger flow based on historical trafic data and current status,and compared with the results of other passenger flow prediction models,the results show that the CNN-BiLSTM passenger flow prediction algorithm has the smallest RMSE and MAE,and relatively high prediction accuracy.

关 键 词：组合模型 CNN-BiLSTM 客流量预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]