Flood Disaster Assessment Method Based on a Stacked Denoising Autoencoder

In recent years, extreme weather has occurred frequently, and the risk of heavy rainfall flooding faced by people risen. It is therefore an urgent requirement to carry out applied research on assessment. We took Henan Province, where a major flood disaster in 2021, as example analyze impact factors urban conduct Indicators were first selected from population, housing, economy, correlation analysis was used optimize indicator system. Then, deep clustering network model based stacked denoising autoencoder (SDAE) constructed, feature information implied indicators abstracted into potential features through coding decoding network, small number express complex relationship between indicators. The results study show that high-risk areas damage Province 2021 account for 2.3%, medium-risk 9.4%, low-risk 80.3%. These evaluation are line with actual situation division grade some more reasonable compared entropy weighting method, which commonly method new does not need calculate weights cope changes conditions. can provide scientific reference management, prevention mitigation, regional planning.