نوع مقاله : مقاله پژوهشی
نویسنده
دانشجوی دکتری،پردیس صنعتی شهدای هویزه، دانشگاه شهید چمران اهواز، اهواز، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسنده [English]
This paper proposes a novel hybrid model, the Fuzzy Deep Kronecker Stacked Autoencoder (FDKSAE), addressing the increasing complexity and dynamism of attacks in Social Internet of Things (SIoT) environments. Initially, raw network data were scaled using Z-score normalization. Subsequently, a deep neural network enhanced with the Gower similarity metric was employed for feature integration. In the detection phase, the deep Kronecker network architecture was combined with a multi-layer autoencoder and fuzzy logic to dynamically react to diverse attack patterns and varying network conditions. For performance evaluation, the N-BaIoT dataset, comprising over seven million samples and ten diverse attack classes, was utilized. Experimental results demonstrated that the proposed model achieved superior performance with simultaneous improvements in accuracy (92%), F1-score (91%), and prediction precision (91%), compared to conventional methods such as GAN, MH CNN AM, TM MLA, and HAD. These achievements underscore that the integration of fuzzy logic with deep architectures can effectively manage uncertainties associated with network traffic and attack patterns.
کلیدواژهها [English]
)