基于元学习和神经网络的快速路瓶颈区识别模型

Recognition Model of Expressway Bottleneck Area Based on Meta-learning and Neural Network

通过较少样本量的快速路瓶颈区数据、高速公路瓶颈区数据、轨道交通瓶颈站点数据,基于元学习方法和神经网络模型,训练得到一个通用的、精度较高的快速路瓶颈区识别模型,该模型可以实时、不间断地主动对快速路段可能会形成瓶颈区的区域和时间进行推理。经实际数据验证,元学习模型准确率为85.3%,召回率为87.1%,F1得分为86.2%。

By using a small sample amount of expressway bottleneck area data,expressway bottleneck area data,and rail transit bottleneck site data,and based on the meta-learning method combined with the neural network model,a general and high-precision expressway bottleneck area identification model is trained.The model can actively and uninterruptedly reason about the area and time that the fast road section may form a bottleneck area.The results show that the accuracy rate of the meta-learning model obtained is 85.3%,the recall rate is 87.1%,and the F1 score is 86.2%.