Machine learning based approach for detecting Distributed Denial of Service attack

  • Author(s) / Creator(s)
  • All elements of the IT industry are expanding, including bandwidth, storage, processing speed. As a result, there are now more cyber threats and attacks, necessitating a creative and predictive security approach that employs cutting-edge technology to combat the danger. The trends will be monitored, and adequate analysis from various sets of data will be utilized to build a model that is based on the information available. Distributed Denial of Service (DDoS) is one of the most prevalent dangers and attacks wreaking havoc on internet-connected computer equipment. This study compares the performance of several machine learning-based classifiers for detecting DDoS assaults before they occur. This experiment made use of data from the benchmark KDD-Cup-1999 DDoS attack. To choose essential characteristics in the context of DDoS detection, I created three distinct types of feature selection techniques. The findings revealed that feature selection approaches can assist domain specialists in understanding the intrusion system’s hidden key patterns and features during DDoS detection. DNN-based deep learning and semi-supervised learning method were also compared with the ML-based classifiers output. The suggested model learns to recognize regular network traffic to detect ICMP, TCP, and UDP DDoS traffic as it arrives. Experiments show that machine learning algorithms may correctly classify the traffic into regular and DDoS. This discovery has long-term implications in a variety of sectors, including national defence, financial institutions, healthcare, and other businesses where sophisticated intrusion detection techniques are required. In the future, I would want to apply similar approaches to a variety of datasets.

  • Date created
  • Subjects / Keywords
  • Type of Item
    Research Material
  • DOI
  • License
    Attribution-NonCommercial 4.0 International