Heavy rainfall super-monomer identification early warning method based on machine learning

The invention relates to the technical field of weather forecast, and particularly discloses a machine learning-based heavy rainfall super-monomer identification and early warning method, which comprises the following steps of: selecting heavy rainfall super-monomer examples, and screening storm characteristic parameters corresponding to the heavy rainfall super-monomer examples to form a characteristic data set; according to the feature data set, a frame based on CatBoost is adopted, and a heavy rainfall super-monomer recognition early warning model is obtained through training; and inputting weather radar networking observation data, topographic data and underlying surface data of a to-be-identified area into the heavy rainfall super-monomer identification early warning model to obtain an identification result, and when the identification result exceeds a set threshold value, judging the occurrence possibility of a heavy rainfall super-monomer. According to the method, multi-source data are fused, and a Cat