Research On Machine Learning-based Intrusion Detection System For Internet Of Things

With the rapid development of the Internet of Things industry,the security of the Internet of Things has been severely challenged.Compared with the characteristics of the traditional Internet,the Internet of Things has the problems of large assets,complex and diverse structures,and lack of computing resources.The traditional intrusion detection system cannot meet the security requirements of the Internet of Things.In response to this situation,this paper applies machine learning algorithms to the intrusion detection system of the Internet of Things to improve detection performance.The specific research work is as follows.1.Apply association rule algorithm to intrusion detection system to mine association rules among all known attacks.The FP-growth algorithm with high efficiency is used to mine the association rules in the intrusion data set,and the intrusion behavior of the Internet of Things is identified through the association rule base.When processing data,the mining algorithm will delete the attack types that do not meet the support requirements,which results in some effective rules being ignored by the algorithm.To solve this problem,this paper divides the intrusion data set according to the type of attack,and then mines each sub-data set in turn.And the improved particle swarm is used to find the most suitable parameter group for each sub-data set,so that the mining algorithm can mine association rules more comprehensively.2.Build an intrusion detection system through deep learning,and build an anomaly detection system based on the CNN-LSTM algorithm.Among them,the convolutional neural network has the characteristics of local receptive field,weight sharing,pooling,etc.;the long-short-term memory neural network can better handle data with long-distance attributes.Therefore,this paper combines these two algorithms to solve the problem that it is difficult to detect new types of attacks by misuse detection,which greatly improves the detection rate of intrusion detection systems.3.Integrate the above two detection systems with each other to design a set of intrusion detection system with integrated algorithm as the core.The two detection modules can be updated with each other,so that the system is automatically upgraded during operation,thereby reducing the maintenance cost of the system.Finally,it is proved by experiments that the Io T intrusion detection system designed in this paper has significantly improved under the four evaluation indicators of accuracy rate,detection rate,accuracy rate,and false alarm rate,which proves the Io T intrusion detection system designed in this paper practicality.