Abnormal data intelligent detection method based on combined machine learning

In this paper, based on abnormal load data, the method of energy Internet load data detection and correction is proposed based on the chaotic directional cuckoo algorithm (COCS) to optimize the Elman-IOC prediction model. Firstly, the Elman-IOC model is trained through a large amount of load data, and the COCS optimization algorithm is selected to find the parameters of the prediction model, the parameter-tuned Elman-IOC model is used to load prediction. And then, starting from the results of load prediction, the abnormal load data in the load curve is detected by error analysis and outliers determination criteria, and the abnormal load data detected is corrected by applying the prediction results.