Research Article
Hybrid LSTM Self-Attention Mechanism Model for Forecasting the Reform of Scientific Research in Morocco
Table 2
Performance evaluation of different models.
| Database | Model | MAE | MSE | MAPE |
| Scopus database | RNN | 0.6900 | 0.0062 | 24.2520 | LSTM | 0.6910 | 0.0078 | 25.9820 | GRU | 0.6710 | 0.0073 | 23.8341 | BiLSTM | 0.6619 | 0.0064 | 22.3741 | Proposed SAM-LSTM | 0.6500 | 0.0062 | 21.9680 |
| Science Direct | RNN | 0.0650 | 0.0067 | 27.398 | LSTM | 0.0680 | 0.0075 | 28.8333 | GRU | 0.0651 | 0.0071 | 27.8921 | BiLSTM | 0.0642 | 0.0067 | 27.4134 | Proposed SAM-LSTM | 0.0630 | 0.0065 | 27.3490 |
|
|