Traffic road condition prediction method and device based on big data

The invention belongs to the technical field of traffic control systems, and particularly relates to a traffic road condition prediction method and device based on big data. The method comprises the following steps: S1, determining whether a road is abnormal or not according to the speed of a passing vehicle on the road; s2, when the road is abnormal, real-time speed information and position information of each vehicle are obtained; s3, determining an abnormal road interval according to the speed information of each vehicle, and constructing a feature vector; step S4, carrying out normalization on the feature vectors; s5, on the basis of a traffic road condition recognition neural network model which is constructed and trained based on big data in advance, the normalized feature vectors serve as model input, model output is obtained, the model output is a triple composed of three probability values, and the three probability values correspond to congestion abnormity, road obstacle abnormity and weather abnorm