Identifying and classifying anomalies in on-line monitoring systems of drainage systems is important to reduce urban water pollution. In the context of big data, the mini-batch K-means combined with the XGBoost drainage network abnormal flow identification and classification model is proposed to...