Development of intrusion detection methods for internet of thingsapplications

With the development of the Internet, IoT devices provide convenience in many fields today. These conveniences increase people's quality of life. People want to remotely monitor and manage the smart city, smart home, and other smart platforms. Remotely monitored and managed applications cause problems such as security problems as well as their advantages. With the increase of IoT platforms, it has become the target point of attackers. Detecting such attacks and preventing security vulnerabilities will increase the rate of use of IoT technology. IoT users are suffered because of network attacks, which are the most preferred by attackers. Besides this suffering, it negatively affects the demand for IoT platforms, as there is access to people's data. In this thesis, attack types such as DDoS, DoS, and Brute Force were examined and methods were developed to detect IoT attacks. Two main contributions were made within the scope of the thesis. First, a laboratory was established using Home Assistant technology to create the smart home platform. In this laboratory environment, 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 IoT devices. Attack types were determined by applying the XGBOOST algorithm to the collected data set. Secondly, the Decision Tree algorithm was applied to the Bot-IoT dataset, which is widely used in the literature. The results were compared by applying other machine learning methods to the data sets used in the thesis.