Research On Feature Engineering In Internet Traffic Classification

Internet traffic analysis and classification technique is an important and a powerful method for network monitoring.It is widely used in various fields,such as intrusion detection system it is like.With the rapid development of the internet,coupled with the enrichment of network applications,the number of internet flow is increasing,which lead to a low classification performance.Currently,the machine learning-based technique has attracted much attention,since it can address the issues that the usage of the dynamic port numbers and the encryption technique at the transport layer in traffic.To address the imbalance issue for Internet traffic data,in this paper,we propose an ensemble feature selection based on balanced partition and min-max strategy.The proposed Min-Max Ensemble Feature Selection(M2-EFS)consists of data partition and Min-Max ensemble strategy.The experimental results demonstrate that the algorithm we proposed can obtain higher performance in most cases,and it could efficiently deal with imbalanced problems.In order to further enhance the performance of internet traffic classification,we extract some features which are based on multiple flows.These features could express the abundant information in the flow,which lead to a high performance.Through the results of the feature selection using M2-EFS,the multiple features more critical than the single flow feature,it could help to enhance the performance in internet traffic classification.In summary,the ensemble feature selection algorithm proposed in this paper could deal with the large scale data and class imbalance problem in traffic classification,and has a certain practical value.