Bamboo Forest Growth Optimization with Deep Learning for Intrusion Detection in IoT-Assisted Wireless Sensor Networks

The Internet of Things (IoT) develops an important research domain and its purpose at interlinking many sensor-enabled devices majorly for data tracking and collecting applications. Wireless Sensor Network (WSN) is a crucial component in the IoT model since its initiation and is developing the most desirable platform for deploying several smart city application areas like disaster managing, intellectual transporting, automation of homes, smart buildings, and other like IoT-based applications. WSNs are exposed to distinct types of security attacks and threats. It is largely because they are extremely restricted in resources like bandwidth, storage, power, and processing power that are utilized in evolving their defence. To make sure their safety, an effectual Intrusion Detection System (IDS) requires that identify these attacks in these constraints. At present, typical IDS are lesser effectual as these malicious attacks can be developed more intelligently, complexly, and frequently. Therefore, this study establishes a Bamboo Forest Growth Optimization with Deep Learning for Intrusion Detection (BFGODL-IDS) in IoT Assisted WSN. The major drivers of the BFGODL-IDS method lie in the efficient detection of the intrusions in the IoT-assisted WSN. For accomplishing this, the BFGODL-IDS method executes a min-max normalization process to scale the input data to a uniform format. Besides, BFGO based Feature Selection (BFGO-FS) model was applied for electing a subset of features. Moreover, an Improved Long Short-Term Memory-Autoencoder (ILSTM-AE) model was used for intrusion detection purposes. Finally, the Bayesian Optimization (BO) model is utilized for the optimum hyperparameter selection process. For validating the higher efficiency of the BFGODL-IDS model, the experimental outcomes are inspected on the benchmark database. The simulation outcome demonstrated the enhanced performance of the BFGODL-IDS method over other DL models.