Scientific early warning method based on buoy monitoring in combination with machine learning

The invention discloses a scientific early warning method based on buoy monitoring and machine learning. The method comprises the following steps: 1, extracting past data from a database for 1-2 years, and carrying out the half-and-half screening of daily data of red tide and daily data of no red tide from the database; the method comprises the following steps: preprocessing historical data, and finding out key factors based on a neural network module in IBM SPSS Modeler; determining a related index range and a threshold value by utilizing an Apriori association algorithm; sending a water quality change warning by utilizing the three conditions; a red tide early warning module is triggered on the premise of meeting water quality change warning, three classification regression methods are screened for classification, and the prediction accuracy is further improved in a bagging mode; the whole process is real-time and automatic, manual intervention is not needed, and monitoring and earlywarning display can be p