Neural network situation prediction method based on dynamic k-means clustering

Abstract

The neural network situation prediction method based on dynamic k-means clustering includes the following steps: 1) Collect basic network security data of a certain system, the data indicators are the number of hosts infected with network viruses, the number of tampered networks, the number of networks implanted with backdoors, The number of security incident reports, the number of counterfeit pages, and the number of security vulnerabilities and high-risk vulnerabilities are normalized for the basic network security data; 2) The normalized basic network security data is clustered using the dynamic k-means clustering algorithm Class, determine the RBF neural network central neuron parameters and the number N; 3) Use the normalized data to participate in RBF neural network training, calculate the RBF neuron width and determine the neuron output; 4) During the training process, the RBF The output weight of the neural network is encoded, and the PSO algorithm is used to obtain the optimal weight, which improves the accuracy of network situation prediction; 5) Use the trained RBF neural network to predict the network situation of a certain month, and compare it with the network situation evaluation value of the month Contrast, calculation error; it has the characteristics of high prediction accuracy.

Description

technical field [0001] The invention belongs to the technical field of network situation prediction, in particular to a neural network situation prediction method based on dynamic k-means clustering. Background technique [0002] The increasingly complex network environment and unpredictable network security vulnerabilities make network management extremely difficult. In recent years, the rise of network situation prediction technology provides convenience for network management by collecting and processing the underlying security element information of the system (such as vulnerability information, the number of hosts infected with viruses, etc.). Network situation prediction technology can not only provide data analysis on the security situation of the future network, but also help network managers make management decisions in advance, so as to improve the network anti-attack capability of the system and reduce the probability of unsafe events. Therefore, network situatio...

AI Technical Summary

Problems solved by technology

The adoption of data mining technology makes the results of network situation assessment and prediction more accurate, but at the same time, due to excessive redundant data and excessive computational complexity, problems such as dimension explosion and slow algorithm operation

Images