Table 1.
Innovations and contributions of the paper.
Fig 1.
Framework of the method.
Fig 2.
The framework of the model.
Table 2.
Dataset sample.
Table 3.
Sequence converted from API to digital number.
Fig 3.
The framework of the transformer.
Fig 4.
Self-Attention with Gate mechanism (SAG) model for encoder.
(a) Self-Attention in Transformer and (b) Self-Attention with Gate mechanism.
Fig 5.
GRU module structure.
Fig 6.
BiGRU module structure.
Table 4.
Different category in the dataset.
Table 5.
Different category in the new dataset.
Table 6.
Confusion matrix.
Table 7.
The configuration of parameter.
Table 8.
Comparison of eight-classification experimental results.
Fig 7.
Accuracy of different algorithms in binary classification on Alibaba Cloud data set.
Table 9.
Comparison of eight-classification with different algorithms on the Alibaba Cloud dataset.
Table 10.
The details of the NSL-KDD dataset.
Table 11.
The details balancing training data after GNGS processing.
Fig 8.
GSB model experimental results on the NSL-KDD.
Table 12.
Comparison of five-classification with different algorithms on the NSL-KDD dataset.
Fig 9.
Accuracy of different algorithms model in five classification on the NSL-KDD dataset.