Table 6.
Comparative analysis of the proposed method with state-of-the-art for hourly data resolution using IHEPC. Bold and underlined results demonstrate the best and runner-up performance.
| References | Method | MSE | RMSE | MAE |
|---|---|---|---|---|
| Mocanu et al. [48] 2016 | FCRBM | - | 0.66 | - |
| Marino et al. [49] 2016 | LSTM-S2S | 0.55 | 0.74 | - |
| Ullah et al. [32] 2019 | MLDB-BLSTM | 0.31 | 0.56 | 0.34 |
| Le et al. [50] 2019 | CNN-BiLSTM | 0.29 | 0.54 | 0.39 |
| Rajabi et al. [52] 2019 | CNN-2D-RP | - | 0.79 | 0.59 |
| Kim and Cho [51] 2019 | SE-AE | 0.38 | - | 0.39 |
| Kim and Cho [55] 2019 | CNN-LSTM | 0.37 | 0.61 | 0.34 |
| Khan et al. [53] 2020 | CNN+LSTM-AE | 0.19 | 0.47 | 0.31 |
| Sajjad et al. [54] 2020 | CNN-GRU | 0.22 | 0.47 | 0.33 |
| Khan et al. [21] 2020 | CNN-MB-GRU | 0.18 | 0.42 | 0.29 |
| Jun and Cho [56] 2020 | CNN-LSTM-MHA | 0.26 | - | - |
| Khan et al. [57] 2021 | AB | 0.56 | 0.75 | 0.67 |
| Proposed | ACL+RGRU | 0.17 | 0.41 | 0.26 |