Attention-based long short-term memory network temperature prediction model

摘要

In order to take full advantage of historical data and improve the accuracy of temperature prediction, This paper combines LSTM with the Attention mechanism and proposes the LSTM-Attenion temperature prediction model.. This model is compared with the traditional LSTM model without attention mechanism. Experimental results show that the attention-based LSTM model is better than the traditional LSTM model without attention mechanism under the RMSE and MAE evaluation indicators. This model has better predictive ability in temperature prediction.