Research On Intrusion Detection Technology Based On Feature Selection And Ensemble Learning

With the rapid development of information technology,it is not only brings convenience to our life,but also brings many security risks.In the era of big data,massive amounts of data generated by individuals using the Internet every day can be stolen by some illegal companies for commercial purposes.For an enterprise,important data is the lifeblood of its development.For a country,the harm caused by the disclosure of important information is immeasurable.Therefore,in order to protect information security and prevent data leakage more effectively,intrusion detection technology emerges at the historic moment.In recent years,the term artificial intelligence is no longer strange,and machine learning has been applied in various fields,including the field of network security.In order to further improve the applicability and timeliness of the intrusion detection algorithm,it is of great significance to study the intrusion detection algorithm based on machine learning.This paper studies intrusion detection technology based on feature selection and integrated learning,which are as follows:Firstly,two intrusion detection data sets are processed.After analyzing the KDD cup99 data set,this paper proposes to use the UNSW_NB15 data set in view of its own defects.Comparing with the former,it shows that the UNSW_NB15 data set reflects the modern network environment more truly.The data which can be directly used for intrusion detection can be obtained by numerical,standardized and normalized experimental processing of the above two data sets.Secondly,the intrusion detection methods and experiments based on feature selection are designed and completed.Aiming at the problem that the data dimension of intrusion detection data set is too high,the method of feature selection is studied deeply.Principal component analysis(PCA)algorithm was used to select the features of the above two data sets,retain the principal component feature attributes and remove the feature columns with low data influence.In the intrusion detection experiment,k-adjacent algorithm and Naive Bayesian algorithm,which are typicalsupervised algorithms in machine learning,are used to classify two kinds of data sets.Experiments show that the feature selection algorithm can improve the detection rate to a certain extent and greatly save the computing time.Finally,the intrusion detection algorithm and experiments based on ensemble learning are designed and completed,and the intrusion detection system model is improved on this basis.The ensemble learning algorithm is added on the basis of feature selection to vote multiple base classifiers,and the classification results are determined according to the voting results.The experiment shows that the accuracy of intrusion detection is improved by adding integrated learning algorithm.The design of the system model is based on the general model,adding feature selection and integrated learning module,which improves the system performance comprehensively.