An exponential smoothing method application to restoring lost data on electric power technical record-keeping in industrial enterprises

Abstract

The paper studies in detail a problem of restoring single data of electric power technical record-keeping systems by means of exponential smoothing. A hypothesis of efficient finding the optimal value of smoothing constant α in the range of 0 < α < 2 is tested. Variation ranges for the mean absolute percentage error of the forecast are specified versus values of the time series variation coefficient for specific electricity consumption processes with different data spread.

Keywords

• exponential smoothing
• lost data restore
• absolute percentage error
• variation