Monitoring method and device for predicting short-time regional electricity consumption condition

The invention provides a monitoring method and device for predicting the short-time regional electricity consumption condition, which are applied to the field of electric power measurement. The method comprises the following steps of respectively setting two prediction periods of regional power consumption as a first preset short-time period and a second preset short-time period, obtaining first preset short-time period power consumption data and second preset short-time period power consumption data, inputting the data into a prediction model, generating a first preset short-time period sample and a second preset short-time period sample, inputting the samples into a trained prediction model, performing time period cooperation operation with a preset long-short-term memory neural network, generating a multi-time scale long-short-term memory neural network, predicting the power consumption condition of a region needing to be predicted by using the multi-time long-short-term memory neural network, and generati