[Retracted] Research on DoS Traffic Detection Model Based on Random Forest and Multilayer Perceptron

Table 5

Accuracy of different MLP structure.

MLP structure		Accuracy
Number of hidden layers	Number of units per hidden layer
CICIDS2017
2	10, 20	0.9978
2	20, 20	0.9982
2	20, 40	0.9978
3	10, 10, 10	0.9983
3	10, 20, 20	0.9966
4	10, 10, 10, 10	0.9921
UNSW-NB15 datasets
2	10, 20	0.7352
2	10, 10	0.9337
2	20, 20	0.9063
3	10, 10, 10	0.9352
3	20, 10, 10	0.9107
4	10, 10, 10, 10	0.8887

According to the comparison of several groups of parameters, we can see that the accuracy rate is the highest when the parameter is (10,10,10), so the parameter (10,10,10) is selected as the optimal parameter of the model and marked in bold in Table 5.