| Autor: |
KOWALSKA-STYCZEŃ, Agnieszka, SOJDA, Adam, WOLNY, Maciej |
| Předmět: |
|
| Zdroj: |
Scientific Papers of Silesian University of Technology. Organization & Management / Zeszyty Naukowe Politechniki Slaskiej. Seria Organizacji i Zarzadzanie; 2023, Issue 177, p325-337, 13p |
| Abstrakt: |
Purpose: The key element of effective electricity management is to improve the accuracy of forecasting its consumption. To create a forecast, data on customers' energy consumption in previous periods is required, and the accuracy of the forecasts depends on the quality and availability of data. The acquired historical data is often incomplete and contains missing values. The aim of the article is therefore to choose an appropriate method of imputation of missing values for one-dimensional time series of energy consumption. Design/methodology/approach: The aim of the article was achieved by using the Known Substring Algorithm (KSSA) to verify the imputation precision. The KSSA algorithm allowed to test of eleven imputation methods, most of which are implemented in the 'ImputeTS' package in R. Based on the RMSE error, the best imputation method was selected for the analyzed series. Findings: As a result of the analyzes carried out, it was shown that the KSSA algorithm is a good tool for selecting the appropriate imputation method in the case of one-dimensional series of electricity consumption series. Based on the RMSE error, ‘auto.arima’ turned out to be the best imputation method for the analyzed objects Research limitations/implications: Future research will concern the use of the KSSA algorithm for a larger number of energy consumption series and with greater variation. Originality/value: The article presents an important problem of the imputation of missing values in the electricity consumption series. Increasing the accuracy of electricity consumption forecasting depends on the quality of the collected data, which are often incomplete and contain missing values. Therefore, the selection of the appropriate imputation method is so important. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
