Effective Intrusion Detection System Using XGBoost

Эффективная система обнаружения вторжений с использованием XGBoost
Abdullah-Al Nahid, Sukhpreet Singh Dhaliwal, Hassan Abbas
2018-06-21

Confusion matrixIntrusion detection systemNSL-KDD datasetNetwork securityXGBoost
As the world is on the verge of venturing into fifth-generation communication technology and embracing concepts such as virtualization and cloudification, the most crucial aspect remains “security”, as more and more data get attached to the internet. This paper reflects a model designed to measure the various parameters of data in a network such as accuracy, precision, confusion matrix, and others. XGBoost is employed on the NSL-KDD (network socket layer-knowledge discovery in databases) dataset to get the desired results. The whole motive is to learn about the integrity of data and have a higher accuracy in the prediction of data. By doing so, the amount of mischievous data floating in a network can be minimized, making the network a secure place to share information. The more secure a network is, the fewer situations where data is hacked or modified. By changing various parameters of the model, future research can be done to get the most out of the data entering and leaving a network. The most important player in the network is data, and getting to know it more closely and precisely is half the work done. Studying data in a network and analyzing the pattern and volume of data leads to the emergence of a solid Intrusion Detection System (IDS), that keeps the network healthy and a safe place to share confidential information.
1
The abstract does not report specific numerical performance results or comparisons with alternative intrusion-detection methods.
2
The model’s parameters can be varied in future research to optimize detection of data entering and leaving networks.
3
The paper applies XGBoost to the NSL-KDD dataset for network intrusion detection.
4
The proposed IDS evaluates network traffic using accuracy, precision, confusion matrices, and other performance parameters.
5
The study aims to improve prediction accuracy and identify malicious network data, supporting network security and data integrity.

network data traffic represented in the NSL-KDD dataset

intrusion-detection classification performance, including accuracy, precision, and confusion-matrix outcomes

Publication Details
Publication Date
2018-06-21
Journal
Publisher
ISSN
Cited by
506
Access Type
Author Information
Authors
Abdullah-Al Nahid
Sukhpreet Singh Dhaliwal
Hassan Abbas
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%