Table 4 Hyperparameter settings for the BERT-based model.
From: A hybrid super learner ensemble for phishing detection on mobile devices
Hyperparameter | Value |
|---|---|
Pretrained model | BERT-base uncased |
Maximum sequence length | 128 |
Batch size | 64 |
Embedding dimension | 768 |
Pooling method | Mean pooling over tokens |
Train-test split | 80–20% |
