The Application Of Neural Network Optimized Based On Adaptive Differential Evolution Algorithm In Intrusion Detection

The development of the Internet is in line with the needs of people.In particular,the mobile Internet,which was born with the combination of mobile communication and the Internet,has developed rapidly.The rapid development of network technology has made the Internet of Things possible.The number of terminals is increasing,and the possibility of being attacked by the network is greatly increased.The problem of network security is becoming more and more prominent.People are more aware of the importance of network security.As the core of network security,intrusion detection technology has become the focus of people's research..With the rapid advancement of technologies such as neural networks and artificial intelligence,intrusion detection technology is also developing in a more intelligent direction.The neural network consists of a large number of parallel neurons,and its structural features are highly fault tolerant,nonlinear,and associative.The use of neural networks in intrusion detection technology has great advantages and occupies an important proportion of intrusion detection.This paper continues to study in the existing neural network model to improve the defects and deficiencies of existing neural networks.Although traditional neural networks have strong nonlinear fitting,they can map various nonlinear relationships.However,the defects are also obvious,such as: easy to fall into local optimum,slow convergence rate,etc.,resulting in high detection false alarm rate and slow speed.For these problems,this paper selects the differential evolution algorithm in biological evolutionary algorithm to optimize it,and constructs an improved adaptive differential evolution neural network model to form a more robust network structure.The improved differential evolution algorithm has the characteristics of global optimization and group intelligence strategy,and is integrated into the neural network algorithm.The weight and threshold of the neural network are taken as the population of the improved differential evolution algorithm,and the error of the neural network is taken as the fitness function of the improveddifferential evolution algorithm.With the weight and threshold of the optimal parameters of the neural network as the target,the proposed improved differential evolution algorithm is used as the optimization function to complete the initial setting of the optimal parameters and realize the optimization of the neural network.This paper applies it to intrusion detection.In the Windows system environment,KDDCUP99 data set(python compiler environment)is selected as the model input for simulation experiments,which realizes the superiority of the algorithm principle.The experimental results show that the optimized neural network model is significantly improved in convergence speed and detection accuracy.It is proved that the adaptive differential evolution algorithm established in this paper optimizes the neural network model for intrusion detection and is feasible and effective.