Deep Pattern Classification Model for Detecting Malicious Network Traffic in IoT-Based Smart Home Environments

The rapid deployment of Internet of Things (IoT) devices in smart home environments increases exposure to malicious network traffic, making accurate and efficient traffic classification essential for protecting connected devices and services. Existing intrusion detection and traffic classification approaches often face challenges in recognizing diverse attack patterns, handling complex traffic characteristics, and maintaining reliable detection performance. To address these limitations, this study proposes a Deep Pattern Classification Model for Detecting Malicious Network Traffic in IoT-Based Smart Home Environments (DPC-MNT-SHE). The proposed model performs network traffic preprocessing, feature representation, deep pattern learning, and multi-class malicious traffic classification using the RT-IoT2022 dataset. DPC-MNT-SHE is built for detecting both normal and malicious traffic patterns and to enhance the discriminative capability over heterogeneous IoT network behaviors. Experimental assessment shows DPC-MNT-SHE achieves 98.5% accuracy, 98.6% precision, 96.5% recall and 98.35% F1-score, which shows its effectiveness in the classification of malicious network traffic over IoT based smart-home environment. The results demonstrate that DPC-MNT-SHE can effectively learn complex traffic patterns and support reliable network security monitoring. Thus, the proposed framework is a practical solution for enhancing malicious traffic detection in IoT-equipped smart homes using deep learning.