Comparação dos algoritmos Random Forest, Random Tree e J48 para detectar ataques DDoS

Abstract

Distributed Denial of Service (DDoS) attack is a type of attack in which a cyber criminal generates a large number of requests for requests to a server, until the server is overwhelmed by so many requests, causing your resources slow down, crash, or slow down your network. There are some techniques and tools to identify this type of attack, such as the signature-based analysis of the SNORT intrusion detection tool. But the problem with this technique is that it does not identify new attacks, only known attacks. To solve this problem, new techniques based on artificial intelligence have emerged that use the learning process to be able to identify and classify through known patterns those traffics that are prone to be malicious, compared to normal traffic. This work analyzes the Random Forest, Random Tree and J48 classification algorithms to verify their effectiveness to identify new DDoS attacks. Simulations are made with different percentages of size for testing and training to verify which one presents the best result.