An Intelligent Approach to Increase the Performance of Threat Detection in IoT
R. Tamilkodi, V. Bala Sankar, Nersu Pavankumar, Potnuru Hemanth Kumar, Uggu Veera Gani Durga, B Pravallika · Advances in computer science research · 2024
The ubiquitous use of IoT (Internet of Things) devices is on the increase.In order for an Internet of Things system to function, it includes all of the necessary hardware, software, networks, sensors, and other parts.The developers of these sensors and devices, however, often omitted details about their minimal resource needs and a slew of security vulnerabilities.In addition, there are a lot of risks associated with the placement of edge networks for IoT devices.The system's performance might be severely compromised by denial-of-service assaults or unlawful sensor hijacking on sites inside the edge network.Our paper presents a model for training and forecasting DDoS attacks using principal component analysis and machine learning methods.The data's dimensionality was reduced using principal component analysis techniques.Metrics for evaluation included precision, accuracy, F1score, and recall.Important parts of the evaluation metrics mentioned earlier are Metrics such as True-Positive, False-Positive, True -Negative, and False -Negative are utilized to assess the impact of the Fourth Industrial Revolution.We used the Training Time to compare each model's training time, which differs from past research.With the use of the CICIDS 2017 and CICIDS 2018 datasets, we assess the performance of our suggested model.In comparison to similar models, the suggested models outperformed them while requiring much less time to train.