A Network Intrusion Detection Method Based on Dynamic Adaptive Clustering Using Inflection Point Radius

Abstract

The network intrusion detection method for realizing dynamic self-adaptive clustering by utilizing the inflection point radius of the present invention is characterized in that it comprises the following steps: performing dimensionality reduction operations on multi-dimensional data records; performing standardized changes on feature vectors of data records after dimensionality reduction; The standardized data record is regarded as a node, and the distance between nodes is calculated by the Euclidean algorithm to represent the degree of correlation between nodes, and the initial clustering is performed according to the degree of correlation between nodes, and the centers of all clusters in the initial clustering are taken as Abstract nodes, repeat the clustering process, and complete the final clustering; judge whether the nodes in each cluster are normal, if the proportion of normal nodes in the cluster is greater than the proportion of abnormal nodes, then classify all nodes in this cluster as normal; All nodes of the cluster are classified as exception class. The method of the invention can realize automatic clustering and optimization of network behavior data, and has obvious advantages in detection rate and stability.

Description

technical field [0001] The invention belongs to the technical field of network security, and relates to a network intrusion detection method for realizing dynamic self-adaptive clustering by utilizing the radius of an inflection point. Background technique [0002] The rapid development of Internet technology has profoundly affected human production and lifestyle, and Internet activities have become an important part of human daily life. However, this high global network penetration also raises a series of security issues. The openness and flexibility of computer network protocols make information systems more vulnerable to intruder attacks, so continuous monitoring and protection of computer networks is required. User authentication, data encryption, and firewalls are some of the traditional techniques used to secure computers. Intrusion Detection Systems (IDS), which emerged later, used specific analysis techniques to detect attacks, identify the source of the attack, an...

AI Technical Summary

Problems solved by technology

However, most of the existing research belongs to the detection algorithm of the supervised type. For network access connections without category labels and data identification features, the intrusion detection model cannot be accurately trained, resulting in low detection accuracy.

Images