Research on Severe Weather Prediction Model Based on Multi-view Learning

In the study of accurate prediction of severe weather, the regional climate characteristics are often not considered for targeted prediction. For the purpose of predicting severe weather more scientifically and accurately, and further accelerate the development of tourism in the Changbai Mountain area, we collect weather data in the Changbai Mountain area, redivision of severe weather based on regional characteristics, and propose Tri-Training severe weather forecasting model based Multi-view Semi-Supervised Learning. The experimental results show that the model proposed in this paper can accurately predict severe weather based on a very small number of labeled weather samples. When the proportion of labeled data samples is 20%, the model accuracy can reach more than 85%, and when the labeled data reaches 80%, the prediction accuracy can reach 95%.