Denetimli makine öğrenimine dayalı ıot ve fiziksel siber güvenlik saldırı tespiti

Internet of Things and cyber physical systems devices are increasingly used in a wide range of applications, from healthcare and wearable devices to critical infrastructures such as nuclear power plants, autonomous vehicles, and smart cities. However, these devices are fundamentally insecure in terms of their programming, electronics, and communications infrastructure. In addition, the increasing number of networks leads to easy security breaches. This represents a wide area of attack that hackers can exploit. Existing methods of detecting unknown system vulnerabilities are very slow, and it is difficult to obtain high accuracy within a short time. An intrusion detection system is the best solution for detecting vulnerabilities and exploits, and can be strategically placed within the network to track traffic from the source to the recipient; the use of machine learning algorithms can also help to give the best level of security. A machine learning system is very accurate, allowing us to instantly predict attacks while reducing the search space. This proposed line of defense reinforces the same spending of security precautions based on said weaknesses due to discerning and motivated cybersecurity practitioners to adopt them in real-world CPS/IoT applications. In this study, the KDD Cup 1999 dataset with 41 available features is used with the implementation of supervised machine learning algorithms such as an artificial neural network, gradient boosting, logistic regression, support vector machine, random forest, decision tree, and Gaussian naïve Bayes as well as BoTNeTIoT-L01-v2. We compare the results in terms of strength, accuracy, and execution time. Our experimental results indicate that the decision tree model has a higher accuracy than the other algorithms.