Next-Generation IoT Intrusion Detection: Leveraging Hybrid Deep Learning Algorithms for Enhanced Security

Wireless sensor node and DDoS (Distributed Denial of Service) is a major attack way that can exhaust the energy of a sensor, disrupt the communication traffic of an IoT network. We must be on the lookout for these attacks. With the quantity of sensor nodes, traditional methods face difficulty in detecting network traffic analysis requiring big data analysis. This paper is dedicated to Network Traffic analysis using Big Data environment and detecting attacks using Deep Learning. Implemented with PySpark in Google Colab, using Keras and Scikit-Learn built-in libraries. To be more specific, training and testing are conducted on the ’CICIoT2023’ / ’TON IoT’ datasets, respectively. When we use a correlation feature reducer, it virtually considers the most significant features for inclusion and reduces all of our features. The first contains a hybrid deep leaning model combining one-dimensional EfficientDet and Gradient Boosting Machine (GBM), comparing it with ten other algorithms. It generated 99.95% binary classification accuracy for the CICIoT2023 dataset and 99.75% for TON IoT data.