@Article{jai.2024.056552,
AUTHOR = {Arkan Kh Shakr Sabonchi},
TITLE = {Optimizing Internet of Things Device Security with a Globalized Firefly Optimization Algorithm for Attack Detection},
JOURNAL = {Journal on Artificial Intelligence},
VOLUME = {6},
YEAR = {2024},
NUMBER = {1},
PAGES = {261--282},
URL = {http://www.techscience.com/jai/v6n1/58395},
ISSN = {2579-003X},
ABSTRACT = {The phenomenal increase in device connectivity is making the signaling and resource-based operational integrity of networks at the node level increasingly prone to distributed denial of service (DDoS) attacks. The current growth rate in the number of Internet of Things (IoT) attacks executed at the time of exchanging data over the Internet represents massive security hazards to IoT devices. In this regard, the present study proposes a new hybrid optimization technique that combines the firefly optimization algorithm with global searches for use in attack detection on IoT devices. We preprocessed two datasets, CICIDS and UNSW-NB15, to remove noise and missing values. The next step is to perform feature extraction using principal component analysis (PCA). Next, we utilize a globalized firefly optimization algorithm (GFOA) to identify and select vectors that indicate low-rate attacks. We finally switch to the Naïve Bayes (NB) classifier at the classification stage to compare it with the traditional extreme gradient boosting classifier in this attack-dimension classifying scenario, demonstrating the superiority of GFOA. The study concludes that the method by GFOA scored outstandingly, with accuracy, precision, and recall levels of 89.76%, 84.7%, and 90.83%, respectively, and an F-measure of 91.11% against the established method that had an F-measure of 64.35%.},
DOI = {10.32604/jai.2024.056552}
}