Intrusions Detection System Using Machine Learning Algorithms

With the development of information technology, networks and the Internet of Things, the number of people using the Internet has increased, as well as the number of internet-connected devices, resulting in important and sensitive data on the Internet, making it vulnerable to penetration. There must be an intrusion detection system, which is one of the safe ways to reduce cyberattacks that the network can be exposed to. The goal of the project is to choose the best algorithm from machine learning algorithms to detect infiltration. In order to achieve the goal, we searched for the latest dataset available on the Internet that meets the need to apply intrusion detection systems and has also not been very largely researched. We got an X-IIOTID data set that applies to the Internet of Industrial Things. It was studied and analyzed using python programming language. It was introduced on several algorithms for machine learning in order to train them to detect and reduce infiltration, and from these algorithms, the best algorithm was (Random Forest). It was the highest accuracy with an accuracy rate of 99.9963% and also the least in terms of false alarms where it reached almost zero compared to other algorithms. It was therefore selected as the best algorithm to be used when designing intrusion detection systems in classifying.