基于人工蜂群算法和XGBoost的网络入侵检测方法研究

为了降低网络入侵检测系统的虚警率,提出一种混合式网络入侵检测方法,将人工蜂群(ABC)算法用于特征提取,XGBoost算法用于特征分类和评价。选择和定义不同的场景和攻击类型,并设计混合式网络拓扑;对预处理后的数据,采用ABC算法进行特征提取,利用XGBoost算法将需要评价的特征进行分类;得到特征的最优子集,利用这些特征完成网络异常检测。在多个公开数据集上的实验结果表明,该混合方法在准确度和检测率方面优于其他方法,且其时间复杂度和空间复杂度较低,表现出较高的检测效率。

NETWORK INTRUSTION DETECTION METHOD BASED ON ABC ALGORITHM AND XGBOOST

To reduce the false alarm rate in network intrusion detection system,a hybrid method is proposed.The artificial bee colony(ABC)algorithm is used for feature extraction,and the XGBoost algorithm is used for feature classification and evaluation.Different scenarios and attack types were selected and defined,and a hybrid computer network topology was designed.ABC algorithm was used to extract features from pre-processed data,and XGBoost algorithm was used to classify the features that need to be evaluated.The optimal subset of features was obtained,and anomaly detection was completed by using these features.The experimental results on several open data sets show that the proposed hybrid method is superior to other methods in accuracy and detection rate.In addition,the time and space complexity of the proposed method is low,which shows a high detection efficiency.