为准确识别城市快速路交通拥堵状态，基于交通状态识别的实时性和变化性，构建了一种基于贝叶斯优化改进的模糊C均值（Fuzzy C-Means, FCM）算法与粒子群优化（Particle Swarm Optimization, PSO）-极度梯度提升树（eXtreme Gradient Boosting, XGBoost）相结合的交通状态识别模型。首先，考虑到FCM算法初始聚类中心的随机性对聚类结果的影响，利用贝叶斯算法对FCM算法进行优化，以避免FCM算法陷入局部最优解；然后，针对XGBoost算法的参数过多且对参数敏感的问题，使用PSO对其重要参数进行优化；最后，采用已分类的交通流数据对PSO-XGBoost模型进行训练，并采用北京市三环快速路的交通流数据对模型进行实例验证和性能对比分析。结果表明，所建模型能较准确和稳定地识别交通状态，识别准确率达99.94%，相比于支持向量机（Support Vector Machine, SVM）模型、随机森林（Random Forest, RF）模型、K-近邻规则（K-Nearest Neighbor, KNN）模型分别提高1.23%, 1.06%, 1.57%，交通状态的识别精度和稳定性都得到有效提升，可为准确提供交通出行信息和缓解交通拥堵提供方法支撑。

In order to accurately identify the traffic congestion state of urban expressway, based on the real-time and variability of traffic state identification, a traffic state identification model based on the combination of the improved Fuzzy C-Means (FCM) algorithm of Bayesian Optimization and the XGBoost algorithm of Particle Swarm Optimization (PSO) was constructed. Firstly, considering the influence of the initial clustering center of the FCM algorithm to the cluster results, Bayesian algorithm was used to optimize the FCM algorithm to avoid the results falling into the local optimal solution. Then, considering the problems of XGBoost algorithm with too many parameters and sensitivity to parameters, some important parameters were optimized using PSO. Finally, the PSO-XGBoost model was trained with the classified traffic flow data, and the traffic flow data of the Third Ring Road in Beijing was used for example validation and performance comparison analysis of the model. The results show that the model proposed can recognize the traffic state more accurately and stably, and the recognition accuracy reaches 99.94%, which is 1.23%, 1.06%, and 1.57% higher than that of the Support Vector Machine (SVM) model, Random Forest (RF) model, and K-Nearest Neighbor (KNN) model, respectively, and the recognition accuracy and stability of the traffic state have been effectively improved, which can provide a method to support the accurate provision of traffic and travel information and the improvement of traffic congestion.