Advances, Systems and Applications
From: VTGAN: hybrid generative adversarial networks for cloud workload prediction
Model | Window size | Training epochs | Time (Sec.) | Precision | Recall | \({F_1 score}\) |
---|---|---|---|---|---|---|
VTGAN (LSTM-based) | 3 | 3000 | 74.2 | 0.826±0.002 | 0.806±0.018 | 0.816±0.009 |
CNN-LSTM | 15 | 235 | 35 | 0.782±0.004 | 0.775±0.017 | 0.778±0.007 |
Stacked LSTM | 15 | 512 | 73.2 | 0.784±0.005 | 0.771±0.003 | 0.778±0.001 |
VTGAN (GRU-based) | 3 | 3000 | 72 | 0.854±0.002 | 0.804±0.003 | 0.828±0.001 |
CNN-GRU | 20 | 235 | 35 | 0.803±0.016 | 0.788±0.017 | 0.795±0.01 |
Stacked GRU | 20 | 255 | 48.9 | 0.743±0.007 | 0.718±0.009 | 0.730±0.008 |