Study on city rainstorm waterlogging warning system based on historical data

Affected by the special geographical environment and climate factors, some cities waterlogging occurred frequently, causing serious economic losses and social impacts. Because of certain topographic factors, once the heavy rain coming suddenly in the city, many of the major streets will be flooded by the water, how to better prevent the occurrence of waterlogging, or predict the depth of waterlogging place, is a very important part of the modern city dealing with natural disasters rainstorm waterlogging. Artificial neural network is a kind of method that human beings use computer to imitate human neuron. It is simulated and predicted by studying sample data. Along with the development of artificial neural network technology, the intelligent prediction based on neural network has been widely used in many aspects. This paper focuses on City rainstorm waterlogging intelligent early warning, BP neural network is used to establish city waterlogging warning system model on a large number of historical data. The prediction results show that the learning model for predictive accuracy is 78%, shortcomings in the model with less training data. Because the rapid economic development in the period of forecast and training years, resulting in some terrain changes greatly, which makes the model prediction results inaccurate.