Akıllı Ev Sistemleri için XGBoost Tabanlı Saldırı Tespit Yöntemi.

In today's smart homes, the infrastructure of IoT (Internet of Things) technology is used. As the use of smart homes increases, cyber attacks in this area are also increasing. It is very important to detect and prevent cyber attacks on smart homes as early as possible. In this study, a machine learning-based method is proposed to detect and prevent cyber attacks against smart homes. First of all, a smart home platform was created using the “Home Assistant” technology. Smart homes make extensive use of “Home Assistant” technology. The created smart home platform makes use of sensors and cameras. People can monitor and manage their homes remotely using sensors and cameras. Seven attacks, namely “brute force ftp”, “brute force ssh”, “dos http flood”, “dos icmp flood”, “dos syn flood”, “syn scan” and “udp scan” were carried out on the developed smart home platform. The collected dataset consists of eight classes with “normal” packages. A total of 435815 sample data were collected for eight classes. XGBOOST algorithm was used on this obtained dataset and attack types were classified. For Hold-out 80:20 and Hold-out 70:30 training test data, 92.55% and 92.49% accuracy were calculated, respectively. The results of the proposed XGBOOST algorithm were compared with the results of other machine learning algorithms and the results were found to be successful. [ABSTRACT FROM AUTHOR]