Research Article
[Retracted] Software Systems Security Vulnerabilities Management by Exploring the Capabilities of Language Models Using NLP
Table 5
Comparison of all the experiments.
| Experiment | Process | Output |
| CNN and FastText embedding | CNN-based processing | Accuracy: 71.89%; precision: 0.88; recall: 0.72; F1-score: 0.77 | Bidirectional LSTM with FastText embedding | Bidirectional GRU or LSTM with global attention | Accuracy: 84.33%; precision: 0.91; recall: 0.84; F1-score: 0.87 | USE model | USE pretrained model with TF 1.0 | Accuracy: 92.61%; precision: 0.95; recall: 0.93; F1-score: 0.93 | NNLM | NNLM-based sentence encoder, with pretrained model | Accuracy: 90.16%; precision: 0.81; recall: 0.90; F1-score: 0.86 | BERT | BERT tokenization and TF Keras modeling | Accuracy: 91.39%; precision: 0.92; recall: 0.91; F1-score: 0.88 | DistilBERT | DistilBERT-based preprocessing of data | Accuracy: 94.77%; precision: 0.95; recall: 0.95; F1-score: 0.94 | BERT | Data preprocessing and tokenization with BERT | Accuracy: 97.44% |
|
|