Integrated classification method and device based on feature reduction and storage medium

Abstract

The embodiment of the invention discloses an integrated classification method and device based on feature reduction and a storage medium, and relates to the field of network security. The method comprises the steps of obtaining multi-dimensional features for an original sample data set; calculating correlation values between the multi-dimensional features and the category labels of the samples; based on the correlation values respectively corresponding to the multi-dimensional features, performing feature reduction on the multi-dimensional features of the sample, deleting a part of features with small correlation values in the multi-dimensional features, and taking the remaining features as a feature reduction result; and based on the feature reduction result of the sample, training a random forest classification model. According to the invention, the efficiency and accuracy of training the intrusion detection model can be improved.

Description

technical field [0001] The invention relates to the field of network security, in particular to an integrated classification method, device and storage medium based on feature reduction. Background technique [0002] With the rapid development of computer network technology, network technology has been widely used in various fields. While computer networks provide convenience and benefits to people, network attacks also pose great challenges to information security. [0003] In order to protect against network attacks, an intrusion detection model can be used to perform intrusion prediction on network flows. In the current process of training the intrusion detection model, due to the high feature dimension of the training samples, the training efficiency of the intrusion detection model is low and takes up a large time cost; at the same time, due to the influence of the distribution of the data set, a single classifier It is easy to form overfitting, which further leads to...

AI Technical Summary

Problems solved by technology

In the current process of training the intrusion detection model, due to the high feature dimension of the training samples, the training efficiency of the intrusion detection model is low and takes up a large time cost; at the same time, due to the influence of the distribution of the data set, a single classifier It is easy to form overfitting, which further leads to low prediction accuracy of the trained intrusion detection model

Images