Abstract:In response to the prevalent issue of missing power measurement data in electric power systems, this study conducted research and practical exploration using statistical methods, interpolation methods, and machine learning methods. Firstly, the paper analyzed the reasons for the missing power measurement data, focusing on the impact of data acquisition, transmission, storage, and other processes on data loss. Subsequently, three methods for data imputation in the presence of missing values were introduced and analyzed in detail, followed by experimental evaluations using various metrics such as correlation coefficient, goodness of fit, and proportion of average absolute error. The experimental results demonstrated that machine learning methods outperformed the other two methods in terms of accuracy and effectiveness in data imputation for missing values. Lastly, the paper summarized the research findings and provided an outlook on the application prospects of machine learning methods in the context of power measurement data imputation. The research outcomes presented in this paper can provide valuable insights for handling missing measurement data in power systems and hold relevance for addressing missing data issues in other domains.