Intrusion Detection Model Based On Random Forests And Improved PSO-CNN

With the rapid development of computer technology and the increasing popularity of computer applications,network security problem has also started to receive much attention.An Intrusion Detection System(IDS)is a system that monitors network traffic for suspicious activity and issues alerts when such activity is discovered.In essence intrusion detection consists of using variables with known values to predict the unknown or future values of other variables.Intrusion Detection has become essential as the use of information technology has become part of our daily lives,and thereby presenting increasing challenges in accurately detecting intrusions.So this paper proposes an intrusion detection model based on the Random Forest and improved PSO optimized CNN structure.Considering the influence of particle parameters on the convergence of the algorithm,a CNN structure optimization algorithm based on Improved PSO is proposed,this paper proposes an improved PSO algorithm-based CNN structure.(LDWPSO-CNN and GQPSO-CNN).LDWPSO-CNN algorithm objective function should be set,and the objective function values of each particle are corresponding fitness values.In every evolution step,the velocity and position for each particle are updated by dynamically tracking its corresponding historical optimal position and the optimal position of swarm population.then get the historical optimal position of particle and the optimal position of swarm population.Update the historical optimal position of particles,the optimal position of swarm population,and the inertial weight.GQPSO-CNN algorithm transforms the CNN structure as particles,which calculates the average value of all particles.then the velocity and position for each particle are updated.Last,the Gaussian Disturbance strategy is used to update the non-optimal particle and improve the global search ability.In order to ensure the effectiveness of the algorithm,this paper used the MNIST data to verify the PSO-CNN algorithm,LDWPSO-CNN algorithm,and GQPSO-CNN algorithm.Experimental results obtained with python demonstrate ability to provide effective optimization and high accuracy according to improved PSO-CNN algorithm.In view of has redundant features of the data set in the process of data processing,and traditional intrusion detection can ' t achieve promising effects,this paper proposes an intrusion detection model and algorithm based on the Random Forest and improved PSO optimized CNN structure(RF-LDWPSO-CNN algorithm,and RF-GQPSO-CNN algorithm).First,introduce Random Forest to data dimension reduction and realize the best compression of data.RF-LDWPSO-CNN algorithm was used to update the inertial weight and improve the global search ability.RF-GQPSO-CNN algorithm was used Gaussian Disturbance to make the population jump out of the local extremum.To improve the convergence speed of the algorithm,the precocity decision was used to take advantage of the elite particles,and optimizes CNN structure with QPSO.This paper introduce Random Forest to data dimension reduction and use these postprocessing data for simulation experiments.The results show that the convergence accuracy of the RF-LDWPSO-CNN algorithm are ahead of CNN algorithm,RF-LDWPSO-BP algorithm,and RF-LDWPSO-SVM algorithm;RF-GQPSO-CNN algorithm is better than traditional RF-GQPSO-BP algorithm,RF-GQPSO-SVM algorithm,RF-GQPSO-DE algorithm,and the accuracy and precision of network intrusion detection of this algorithm is increased,the rate of false positives is reduced.