Detection of Cyber Attacks in Network Using ML

Abstract

Cybercrime is on the rise everywhere and takes advantage of different flaws in the computing environment. Paying ethical hackers boosted their focus on finding flaws and suggesting solutions. Due to machine learning's success in solving problems related to cyber security, it has recently become a topic of significant relevance. Major concerns in cyber security, such as intrusion detection, malware categorization, and detection, have been addressed using machine learning approaches. Although though it cannot fully automate a cyber security system, machine learning may be able to identify cyber security threats more effectively than other software-oriented approaches, which lessens the stress on security analysts. In this research, we suggest using a machine learning model to identify the network attack. To obtain reliable predictions, some machine learning approaches, such as Random Forest, SVM, Gradient Boosting have been applied. The dataset CSE-CIC-IDS2018 was used to train the model. As a result, efficient adaptive techniques, including different machine learning algorithms, can increase detection rates. The basic objective is to ascertain whether the network is being attacked. Random Forest gave highest accuracy of 99.99%.