An Enhanced Intrusion Detection System Model Using Machine Learning Algorithm

Intrusion Detection Systems (IDS) play a pivotal role in safeguarding the integrity and performance of an organization. Throughout recent years, various approaches have been devised and put into practice to fortify the security, reliability, and availability of computer networks within enterprises. Our research specifically on IDSs constructed using Machine Learning (ML) techniques. In our research, we have conducted an in-depth analysis of the CICIDS2017 intrusion detection dataset, which serves as the foundation for training and evaluating our machine learning models. We have implemented several machine learning approaches, including Decision Trees (DT), Extra Trees, Random Forest, and XGBoost, to study their effectiveness in detecting intrusions and ensuring the security of computer systems. In our experiments, we explored the use of Principal Component Analysis (PCA) for feature selection and HyperOpt for optimization. Our findings revealed that the XGBoost classifier achieved the highest accuracy among all three other methods (DT,ET,RF) the methods, with an impressive accuracy rate of 99.331%. This suggests that XGBoost is particularly effective in our context.