Intrusion Detection System with SVM and Ensemble Learning Algorithms

• Published in: SN computer science 2023-09, Vol.4 (5), p.517, Article 517
• Main Authors: Johnson Singh, Khundrakpam, Maisnam, Debabrata, Chanu, Usham Sanjota
• Format: Article
• Language: English
• Subjects: Accuracy; Algorithms; Artificial intelligence; Classification; Clustering; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data Structures and Information Theory; Datasets; Decision trees; Denial of service attacks; Ensemble learning; False alarms; Feature selection; Genetic algorithms; Information Systems and Communication Service; Intrusion detection systems; Machine learning; Original Research; Pattern Recognition and Graphics; Prediction models; Research Trends in Computational Intelligence; Software Engineering/Programming and Operating Systems; Support vector machines; Vision
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-023-01954-3

One of the most effective methods of training a model for intrusion detection requires a very good selection of features from the data and efficient and robust training algorithms to facilitate a better prediction model. Choosing features scoring above a certain threshold allows for the removal of unrelated features following which makes the job of a prediction model easier. The study aims to identify and select the highly correlated features after feature reduction for training the model and then employ various machine learning algorithms to make the classifications with tree-based ensemble learning techniques and non-linear SVM. The dataset from NSL-KDD which is a version derived from the KDD’99 Cup dataset is considered. Implementation is carried out in Python 3 using the scikit-learn machine learning library which is built upon SciPy. Further, the performances of various machine learning classifiers will be evaluated to test for and compare the detection metrics.