Intrusion detection method based on long-short-term memory self-encoding classifier under Internet of Things

Abstract

The invention discloses an intrusion detection method based on a long-short-term memory self-encoding classifier under the Internet of Things. Deep learning is an effective mode for realizing intrusion detection, but the detection capability of a traditional simple and unitary deep learning model, such as CNN and RNN, is very limited for increasingly complex network data and network attacks. Therefore, in order to further improve the detection precision and efficiency of the model, an unsupervised learning auto-encoder is added, nonlinear dimensionality reduction is performed on data by using strong feature extraction capability of the unsupervised learning auto-encoder, then a long and short term memory model and the auto-encoder are combined, and the characteristics of the two models based on time series and nonlinear dimensionality reduction are fully combined to detect the network traffic. Comparison tests prove that the long-short-term memory self-encoding classifier is superior to BGRU, BLSTM and gating cycle unit self-encoding classifiers in detection precision.

Description

technical field [0001] The invention belongs to the field of intrusion detection and deep learning, and in particular relates to an intrusion detection method based on a long-short-term memory self-encoding classifier under the environment of the Internet of Things. Background technique [0002] The Internet of Things connects things and things, people and things through various information collection devices and the Internet, so as to manage and control things more intelligently. With the advent of a new generation of information technology, the Internet of Things has developed rapidly and is now widely used in communications, medical care, education, industry, and agriculture. A large number of IoT devices connect all things together and promote the development of human society. However, IoT devices are extremely vulnerable to cyber attacks from hackers. For example, IoT devices have no protection measures against network attacks because of their simple performance, and ...

AI Technical Summary

Problems solved by technology

As a network barrier, intrusion detection can detect malicious traffic very well. However, with the increasing variety of network attack methods and the sharp increase in network traffic, the detectors of the intrusion detection system need to be constantly updated, which also affects the detection rate of the intrusion detection system. and false positive rate put forward higher requirements

Images