Research Article
[Retracted] Software Systems Security Vulnerabilities Management by Exploring the Capabilities of Language Models Using NLP
| Input: security- and nonsecurity-related text with labeling | | Process: | (1) | Data preprocessing and tokenization to create a BERT layer: | | FullTokenizer = bert.bert_tokenization | | FullTokenizer | | bert_layer = hub.KerasLayer | | (“https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/1,” | | trainable = False) | (2) | For data set creation purposes, padding of the data batches to be done, to bring all the training sequence to a consistent length | (3) | Test and train data set batches are created | (4) | DCNN model building as per the specifications provided in the tables above | (5) | Training the model with the specifications provided in Table 2 | (6) | DCNN model compilation: DCNN(tf.keras.Model) DCNN = DCNN (vocab_size, emb_dim, nb_filters, FFN_units,nb_classes, | | dropout_rate) | (7) | Fit the model with training data | (8) | Model evaluation with test data | | Output: | | Accuracy: 97.44% |
|