An Intelligent Imputation Method for Electricity Data

The missing daily frozen forward active energy data in electricity energy data acquisition system will affect electricity analysis and electricity charge calculations. To solve the problem, this paper proposes an intelligent electricity data imputation method based on electricity data on same period and average electricity data in the same industries. To find the date on same period, matching algorithm of similar date and distance function of dates are proposed. The results show that during Spring Festival, May Day, National Day and non-holidays, for the two subdivisions of no daily frozen data at the end and daily frozen data at the end, the intelligent electricity data imputation method is better than the transaction imputation method and arithmetic average method. Besides, the standard error of intelligent electricity data imputation method is reduced by 34.43% and 28.25% compared to the transaction imputation method and the arithmetic average method respectively.