Research on energy big data cleaning based on multi-source data analysis

In view of the difficulties in extracting a unified anomaly detection mode and the low continuity and accuracy of abnormal data correction in the process of energy big data cleaning, a research method of energy big data cleaning based on multisource data analysis is proposed. Firstly, the normal clusters are obtained based on the improved multi-source data analysis to realize the classification feature recognition of energy big data. According to the recognition results, the boundary sample acquisition method of normal clusters is realized, the energy data anomaly detection algorithm is optimized, and the energy big data cleaning model is constructed. Finally, the reliability of this method is verified by experimental analysis.