Power prediction method, system and equipment based on group algorithm

The invention is suitable for the technical field of electric power prediction, and provides an electric power prediction method, system and device based on a swarm algorithm, and the method comprises the steps: S1, obtaining electric power prediction data through an API interface, carrying out the data cleaning and standardization of the obtained prediction data, and converting the data into a data set which is consistent in structure and is suitable for analysis; s2, combining PSO (Particle Swarm Optimization) with a genetic algorithm by adopting a hybrid swarm intelligence algorithm, and optimizing a deep learning model; s3, optimizing the combination weight of the plurality of basic prediction models by using a swarm intelligence algorithm to form an integrated model; s4, performing performance prediction evaluation on the integrated model; s5, deploying the integrated model into a real-time prediction system, continuously receiving new data, performing prediction, and establishing a feedback mechanism; a