Abstract: Due to the advancement of wireless communication, there are several online security risks. The importance of intrusion detection systems (IDS) in computer and network security cannot be overstated. The experiment dataset in this research was the KDDCUP'99 (Knowledge Discovery Dataset) intrusion detection dataset. Due to intrusion detection's fundamental properties, there is still a significant imbalance between the classifications in the dataset, which makes it more difficult to apply machine learning to intrusion detection efficiently. IDS techniques come in a wide variety and yield results with varying degrees of precision. This calls for the creation of an efficient and reliable intrusion detection system. In this paper, a method for creating effective IDS that makes use of the random forest classification algorithm and principal component analysis (PCA) is proposed. While Random Forest (RF), an ensemble classifier that outperforms other standard classifiers for the accurate classification of attacks, PCA will assist in organizing the information by reducing its dimensionality. Together with the Confusion Matrix, a performance evaluation tool, we have also employed other approaches for model evaluation and selection, including as accuracy, precision, recall, and f-score.
Keywords: IDS, Knowledge Discovery Dataset, PCA, Random Forest
|
DOI:
10.17148/IJARCCE.2022.116100
