ALDO CANI, AURORA SIMONI
KEYWORDS : Cyber Attacks, machine learning, attack detection, classification, object identification, cyber security.
Abstract
Cyberattacks are already a threat that continues to grow every day for institutions, universities, organizations or even simple Internet users. Attacks that succeed often cause data deletion, encryption and denial of access causing huge losses which in many cases are unrecoverable. Intrusion Detection System (IDS) technologies can help to detect attacks but in the case of sophisticated cyberattacks, they do not give the expected results because these types of attacks require new technologies mainly based on artificial intelligence. A promising method for preventing attacks is the Machine Learning (ML) method, which has achieved very good results for identifying and preventing cyber security threats. These methods improve cyber security based on data traffic analysis. To achieve high results, ML algorithms must be trained with large datasets in order to identify patterns of cyberattack behavior. This process enables them to detect attacks in real time, helping to prevent the possible consequences of the attack. In this paper, we will discuss the attacks and their types, and we will also perform machine learning identification when these attacks occur. We will analyze their detection methods as well as analyze the best algorithms for detecting attacks in real time. The study is a valuable resource for those who wish to deeply understand these attacks and their detection through machine learning.