Prediction of urban water accumulation points and water accumulation process based on machine learning

With the development of urbanization, global warming, rain island effect and other factors, cities around the world are facing more frequent and intense flood events. In order to deal with the damage caused by urban flood effectively, it is increasingly important to accurately predict and characterize the information of the flood in cities. In recent years, the rise of machine learning methods provides a new technical means for flood prediction. In this study, Naive Bayes (NB) and Random Forest (RF) algorithm were used to forecast the waterlogging point and the waterlogging process at the waterlogging point respectively to achieve the goal of predicting the whole process of urban waterlogging. Compared with the actual result, the four evaluation indexes (P, R, A and F-1) of the NB classification models are 91%, 90.5%, 98.9% and 90.7% respectively, and the three regression indexes (MAE, MRER and RMSE) of the RF regression model were respectively 0.95%, 9.53% and 1.21%. The results demonstrated that the prediction result of NB model for waterlogging point is reliable, and the process of waterlogging predicted by RF model is also consistent with the actual situation, which verify the validity and applicability of the NB model and RF model. This research is expected to provide scientific guidance and theoretical support for urban flood disaster mitigation and relief work.