Fig. 3
From: A Large-Scale Residential Load Dataset in a Southern Province of China
Comparison of various imputation methods for addressing missing data. The selected forward-backward imputation method is denoted by the black line, while the shaded background indicates the presence of missing data. Three days in June 2022 are selected. Additional imputation methods, including linear interpolation, forward imputation, and backward imputation, are also evaluated for comparison purposes. Forward imputation addresses missing values by utilizing data from the preceding week, whereas backward imputation employs data from the subsequent week to fill in gaps. The forward-backward imputation method integrates both approaches by averaging the data from the week prior and the week following the missing values.
