Urban Temperature Prediction Model Based on CNN-Bi-LSTM Model

With the frequent occurrence of extreme weather caused by increased urban pollution, accurate prediction of urban temperature can predict the time of extreme weather in advance to a certain extent and take preventive measures to reduce unnecessary casualties. The research topic of this paper is to predict urban temperature by analyzing the relationship between urban environmental variables and urban temperature. The research method of this paper is to establish a CNN-Bi-LSTM model based on deep learning, and to compare and analyze the prediction effect of CNN-LSTM, CNN-Bi-LSTM and multiple linear regression model using RMSE as the evaluation standard. The results show that the RMSE of the model used in this paper is 0.246 and 0.06 lower than the CNN-LSTM model and 1.073 lower than the multivariate linear regression model on the training and test sets, respectively. Meanwhile, this model has a better prediction effect at extreme temperatures. The results show that this model has high accuracy in both roots mean square error (RMSE) and urban temperature prediction.