Parallel Intrusion Detection Method and System Based on Imbalanced Data Deep Belief Network

Abstract

The invention discloses a parallel intrusion detection method based on unbalanced data deep belief network, which reads unbalanced data set data, uses an improved NCL algorithm to perform under-sampling processing on unbalanced data, reduces the proportion of most samples, and makes The data distribution of the data set is balanced; the improved differential evolution algorithm is used to optimize the parameters of the deep belief network model on the distributed memory computing platform Spark platform to obtain the optimal model parameters; the feature extraction of the data set data is performed, and then the weighted The kernel extreme learning machine is used for intrusion detection and classification. Finally, multiple weighted kernel extreme learning machines with different structures are trained in parallel by multi-threading as the base classifier, and a multi-classifier intrusion detection model based on adaptive weighted voting is established for parallel intrusion. detection. The invention can solve the technical problems that the existing intrusion detection method lacks pertinence to the unbalanced data set and the training time is too long, and improves the speed of optimizing the parameters of the deep belief network model.

Description

technical field [0001] The invention belongs to the technical field of intrusion detection, and more particularly, relates to a parallel intrusion detection method and system based on unbalanced data deep belief network. Background technique [0002] With the development of society, the issue of network security has been paid more and more attention by people. Intrusion detection method is an effective and active defense method for network security problems. It judges whether there is abnormal intrusion behavior in the network by detecting information such as traffic in the network. Compared with the firewall, the security of the intrusion detection method is better. It not only requires less resources, basically does not affect the normal operation of the system, but also can be adjusted dynamically. [0003] The current mainstream intrusion detection methods mainly include: 1. Intrusion detection methods based on unbalanced data. The intrusion detection methods are mainly...

AI Technical Summary

Problems solved by technology

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a parallel intrusion detection method and system based on unbalanced data deep belief network, the purpose of which is to solve the lack of pertinence of existing intrusion detection methods for unbalanced data sets At the same time, the speed of optimizing the parameters of the deep belief network model is improved. Finally, the invented method can effectively improve the detection accuracy and detection speed of intrusion detection.

Images