The article examines the categories of machine learning models: ensemble methods implemented by the random forest algorithm and boosting (XGB classifier, XGB regressor); linear models (logistic regression); classifier based on deep neural networks. The results of a study of machine learning models are presented, where the “Random Forest” model obtained the best results. Graphs of the ROC curve for each considered machine learning model are constructed. Various parameter values are considered for the selected model. The best model parameters have been selected.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Investigation of Machine Learning Models for Detecting Network Anomalies

  • Svetlana Govorova,
  • Sergey Melnikov,
  • Egor Govorov,
  • Mohammad Shahid

摘要

The article examines the categories of machine learning models: ensemble methods implemented by the random forest algorithm and boosting (XGB classifier, XGB regressor); linear models (logistic regression); classifier based on deep neural networks. The results of a study of machine learning models are presented, where the “Random Forest” model obtained the best results. Graphs of the ROC curve for each considered machine learning model are constructed. Various parameter values are considered for the selected model. The best model parameters have been selected.