Research On The Generation Mechanism Of The Detector In The Real-valued Negative Selection Algorithm

The negative selection algorithm is an intelligent algorithm that draws on the biological immune system and is one of the important algorithms of the network anomaly detection algorithm.The real-valued negative selection algorithm expands the problem-solving space from a simple single expression to a flexible expression for the first time,which is more in line with real practical application problems.The flexible change real-valued negative selection algorithm changes the constant detector size,greatly improves the coverage efficiency of the detector for non-self-regions,and reduces the number of detectors generated by the negative selection algorithm.How to generate a valid detector is a key issue in the selection algorithm.Although the real-valued negative selection algorithm of the variable radius detector has been greatly improved in the detection rate and the false alarm rate,for large and complex practical engineering problems,the redundant detector causes the algorithm to be detected.The phase takes a lot of time,and the parameters of the existing negative selection process in the detector generation phase are selected empirically,which makes the detection performance of various algorithms different.Based on this problem,this paper proposes a new negative selection algorithm based on hierarchical clustering of detector sets.After generating the detector set,it performs hierarchical clustering from top to bottom to obtain the detector cluster center set.The main purpose is to reduce the number of detector sets,thereby reducing the computational cost of the test phase,while improving the detection rate of the algorithm and reducing the false alarm rate.The detector's own radius has a direct impact on the detection performance.As one of the basic algorithms widely used in the negative selection algorithm,V-detector is of great significance for the subsequent research improvement in this field.The authors will use a new set of auto-parameter evaluation methods to calculate the optimal auto-radius of the V-detector through an iterative algorithm by analyzing the auto-radius of the V-detector.After generating the variable radius detector with different parameters,the optimal self-radius of the training set is obtained through experiments,which provides a better data set for subsequent research,thereby improving the coverage area of the detector for the non-self area.