Work

Effective Machine Learning Classifiers for Intrusion Detection in Computer Network

Year: 2023
Type: article
Abstract: Cyber security has finally become inevitable due to the increase in use of internet and computer networks. This has given access to cyber-attacks in the network services. However, Intrusion detection systems (IDSs) have been incorporated into networks so as to overcome these huge challenges. IDSs are capable of identifying malicious or abnormal activity in the network and draw the attention of the network administrator to it. Furthermore, approaches based on machine learning (ML) are able to increase IDS effectiveness. In this study, the NSL-KDD CUP dataset was used to develop and validate three individual models using three supervised machine learning algorithms: Classification and Regression Tree (CART), Multinomial Logistics Regression (MLR), and K-Nearest Neighbor (KNN).Data preprocessing and Feature Selection was initiated in order to remove outliers and imbalance in the dataset and optimally select the best feature to avoid data redundancy. Performance evaluation was conducted on each of the model developed with the following metrics as training time, precision accuracy, sensitivity, , f-Score and specificity, The evaluation`s final findings indicate that KNN is the most effective classifier, with a classification accuracy of 99.38% and a training time of 13.64 seconds with the lowest error rate (0.006161), making it the best model and encouraging further study.
Language: English
Open Access status: diamond
OpenAlex ID: W7217867404
FWCI: 0.388
Cites: 14
Cited by: 5
Related to: 10