Multi-model fusion intelligent power grid cloud data center resource load prediction method

The invention provides a multi-model fusion smart grid cloud data center resource load prediction method, which comprises the following steps: S1, firstly establishing an ARIMA prediction model, a BPNN prediction model and an LSTM prediction model, and predicting a time sequence and future workload by using the established models; s2, using a CRITIC method in an objective weighting method to carry out weighted combination on prediction results of the three prediction models and prediction results of multiple models through the CRITIC objective weighting method to obtain a combined prediction result, and predicting a future error according to a real error; and S3, performing error correction on the combined prediction result to obtain a final prediction result. The prediction model and method provided by the invention have good prediction precision and certain generalization ability, can accurately predict the change trend of the load of the cloud data center, and can effectively improve the network resource u