| Title |
The Interpolation Method of Electricity Consumption Data by Business Using LSTM and GRU |
| Authors |
장민영(Min-young Jang) ; 이정일(Jung-il Lee) ; 정남준(Nam-joon Jung) ; 고영준(Yeong-jun Koh) |
| DOI |
https://doi.org/10.5370/KIEE.2023.72.3.413 |
| Keywords |
Interpolation Method; Electricity Consumption Data; LSTM(Long Short-Term Memory); GRU(Gated Recurrent Unit) |
| Abstract |
The electricity consumption data is collected using the communication network from the smart meter. Missing values occur depending on the status of the meter and the communication network environment. In previous studies, missing value interpolation was suggested as a method of using electricity users' past electricity usage patterns, but autocorrelation in artificial intelligence techniques can reduce generalization performance. In this study, the method of generating deep learning models by business improved the stability and efficiency of the model compared to the method of interpolation by electricity user. An interpolation model creation method was presented by comparing four LSTM models with higher accuracy than GRU. |