A Hybrid SVM-LSTM Temperature Prediction Model Based on Empirical Mode Decomposition and Residual Prediction

Weather prediction is one of the hot topics in artificial intelligence. In this paper, three new temperature prediction models based on historical data are proposed for two important meteorological indexes, the maximum temperature and the minimum temperature. The first model is to construct SVM model to predict the residual error of LSTM model, then add the prediction results of the two models to get the final prediction result. The second model is to use empirical mode decomposition (EMD) to decompose the original data, then use the combination forecasting model to predict the subsequences, and finally summarize the prediction results. The third model is to combine the advantages of the first and second models. First, EMD is used to decompose the original sequence. Then, the first model is used to predict each subsequence. Finally, the predicted values of all subsequences are superimposed to obtain the final predicted value. Based on the temperature data of Washington and Los Angeles, the three models are tested and analyzed in this paper. The experimental results show that the third model proposed in this paper, which is based on EMD and residual prediction SVM-LSTM model, has better prediction accuracy than other models.