基于时空信息融合的高速公路区域货运量预测模型

赵利英,王占中

Prediction model for regional freight volume on highways based on spatiotemporal information fusion

Liying ZHAO,Zhanzhong WANG

表 1 不同长短期记忆网络模型的货运量预测结果对比

Tab.1 Comparison of freight volume prediction results among different long short-term memory models

gi 日期 NFVT T-LSTM S-LSTM TS-LSTM NFVP er/% NFVP er/% NFVP er/% 长春市 2020–11–08 −1.760 35 −0.425 58 −75.82 −1.57041 −10.79 −1.61972 −7.99 2021–01–09 −1.041 60 −0.987 70 −5.17 −0.84213 −19.15 −1.02464 −1.63 2021–05–13 1.500 42 1.092 39 −27.19 1.62375 8.22 1.57568 5.02 吉林市 2020–11–09 −1.013 21 −0.965 39 −4.72 −0.58918 −41.85 −0.99375 −1.92 2020–12–15 0.072 36 0.026 48 −63.41 0.087948 21.53 0.08562 18.33 2021–01–07 −0.542 42 −0.606 17 11.75 −0.72326 33.34 −0.58881 8.55 松原市 2020–11–14 0.562 48 0.408 93 −27.30 0.53919 −4.14 0.55495 −1.34 2021–01–27 −2.074 66 −2.126 46 2.50 −2.77361 33.69 −2.06008 −0.70 2021–03–28 1.101 01 1.129 07 2.55 1.02592 −6.82 1.09384 −0.65