HPSO and multi-core LSSVM based network intrusion detection

Summary:

Aiming at the problems of low accuracy and insufficient feature extraction ability of existing intrusion detection technology, a network intrusion detection model based on hybrid particle swarm optimization and multi-core least squares support vector machine was constructed.Firstly, aiming at the problems of weak generalization ability and poor learning ability of single-kernel least squares support vector machine, this model combines the advantages of polynomial kernel function and radial basis function, and constructs multi-kernel least squares support vector machine;A hybrid particle swarm algorithm for intrusion detection data feature extraction and multi-core least squares support vector machine parameter optimization; finally, the extracted features are used as the input of the multi-core least squares support vector machine after parameter optimization to realize the intrusion detection data set.classification identification.To evaluate the effectiveness of the proposed model, experiments are conducted based on NSL-KDD, UNSW-NB15 and CICIDS-2017 datasets, and the experimental results show that the detection effect of the proposed model is significantly better than other traditional models.

Content Directory:

1 Multicore LSSVM

2 Hybrid particle swarm algorithm HPSO

2.1 Particle Swarm Optimization

2.2 Binary Particle Swarm Optimization

2.3 Hybrid particle swarm algorithm HPSO

2.3.1 Particle coding

2.3.2 Fitness value function based on MKLSSVM

2.3.3 HPSO algorithm process

3 Network intrusion detection model based on HPSO and multi-core LSSVM

4 Simulation experiment

4.1 Experimental environment and evaluation indicators

4.2 Experimental data

4.3 Model parameter settings

4.4 Evaluation index of experimental results

4.4.1 Comparison of LSSVM models under different kernel functions

4.4.2 Comparison of HPSO-MKLSSVM with LSSVM and MKLSSVM

4.4.3 Comparison of HPSO-MKLSSVM with PSO-MKLSSVM and BPSO-MKLSSVM

4.4.4 Comparison with other existing detection models

5 Conclusion

With the development of information technologyÿ