Comparing the Performance of Adaptive Boosted Classifiers in Anomaly based Intrusion Detection System for Networks
Title | Comparing the Performance of Adaptive Boosted Classifiers in Anomaly based Intrusion Detection System for Networks |
Publication Type | Conference Paper |
Year of Publication | 2019 |
Authors | Sivanantham, S., Abirami, R., Gowsalya, R. |
Conference Name | 2019 International Conference on Vision Towards Emerging Trends in Communication and Networking (ViTECoN) |
Keywords | adaptive boosted classifiers, boosting, boosting process, boosting techniques, composability, computer network security, Correlation, Correlation based feature weighted Naïve Bayes, data mining, Decision trees, detection rate, discretization, hybrid anomaly based intrusion detection system, Intrusion detection, intrusion detection Kaggle dataset, intrusion detection system, Intrusion Detection Systems, learning (artificial intelligence), lengthy detection process, Metrics, naïve Bayes classifier, network based intrusion detection, network intrusion detection, pattern classification, pubcrawl, Random Tree, Resiliency, Tools, Training |
Abstract | The computer network is used by billions of people worldwide for variety of purposes. This has made the security increasingly important in networks. It is essential to use Intrusion Detection Systems (IDS) and devices whose main function is to detect anomalies in networks. Mostly all the intrusion detection approaches focuses on the issues of boosting techniques since results are inaccurate and results in lengthy detection process. The major pitfall in network based intrusion detection is the wide-ranging volume of data gathered from the network. In this paper, we put forward a hybrid anomaly based intrusion detection system which uses Classification and Boosting technique. The Paper is organized in such a way it compares the performance three different Classifiers along with boosting. Boosting process maximizes classification accuracy. Results of proposed scheme will analyzed over different datasets like Intrusion Detection Kaggle Dataset and NSL KDD. Out of vast analysis it is found Random tree provides best average Accuracy rate of around 99.98%, Detection rate of 98.79% and a minimum False Alarm rate. |
DOI | 10.1109/ViTECoN.2019.8899368 |
Citation Key | sivanantham_comparing_2019 |
- intrusion detection Kaggle dataset
- Training
- tools
- Resiliency
- Random Tree
- pubcrawl
- pattern classification
- network intrusion detection
- network based intrusion detection
- naïve Bayes classifier
- Metrics
- lengthy detection process
- learning (artificial intelligence)
- Intrusion Detection Systems
- intrusion detection system
- adaptive boosted classifiers
- Intrusion Detection
- hybrid anomaly based intrusion detection system
- discretization
- detection rate
- Decision trees
- Data mining
- Correlation based feature weighted Naïve Bayes
- Correlation
- computer network security
- composability
- boosting techniques
- boosting process
- boosting