Table 4 AUC of SISA variants in different datasets.
From: Session interest model for CTR prediction based on self-attention mechanism
Datasets | FNN | AVG | MAX | ATT | |||
|---|---|---|---|---|---|---|---|
SN | IN | SN | IN | SN | IN | ||
Avazu | 0.7825 | 0.7841 | 0.7937 | 0.7902 | 0.8016 | 0.8083 | 0.8324 |
Criteo | 0.7758 | 0.7702 | 0.7814 | 0.7843 | 0.7938 | 0.8173 | 0.8244 |
Books | 0.8026 | 0.8272 | 0.8395 | 0.8331 | 0.8405 | 0.8501 | 0.8965 |
Electronics | 0.7938 | 0.8198 | 0.8306 | 0.8274 | 0.8317 | 0.8457 | 0.8772 |
