Fuzzy deep learning method for urban traffic cognition under sudden epidemic disease

The invention discloses a fuzzy deep learning method for urban traffic cognition under sudden epidemic diseases. A fuzzy theory and a graph neural network are fused to form a fuzzy deep learning framework so as to deal with complex spatial-temporal characteristics of traffic cognition related data and uncertainty of external data, and finally urban traffic flow prediction is performed. According to the framework, a fuzzy reasoning mechanism is utilized to obtain fuzzy representation of sudden epidemic situation related data and extract corresponding features so as to process irregularity of urban traffic external factors; meanwhile, traffic cognition related data are processed through a graph neural network, so that the dynamic space-time correlation of traffic data is captured; and finally, a traffic flow prediction result of the model is applied to urban traffic cognition work. 本发明公开面向突发流行病下城市交通认知的模糊深度学习方法。将模糊理论和图神经网络进行融合，形成一种模糊深度学习框架，以处理交通认知相关数据的复杂时空特征和外部数据的不确定性,最终进行城市交通流量预测。本框架利用模糊推理机制获得突发流行病疫情相关数据的模糊表示并提