Multivariable time series data prediction model based on double-window mechanism

The invention discloses a multivariable time series data prediction model based on a double-window mechanism. The model comprises a short sequence processing module and a long sequence processing module. The short sequence processing module is used for processing short time sequence data and extracting stable recent information from the short time sequence data; and the long sequence processing module is used for processing the long-time sequence data and extracting periodic and seasonal long-term information from the long-time sequence data. And finally, the two parts are combined, so that recent information and a long-term rule are combined, and a better prediction result is obtained. According to the method, the long-time sequence data is effectively utilized, so that the time sequence prediction accuracy is improved. The system provided by the invention obtains the best RMSE and MAE in all data sets. Therefore, the model is superior to a model which purely uses the short sequence due to the fact that the