Nature-Inspired Algorithms for Identifying and Detecting Criminal Activities in the Internet of Things: A Review of Methods and Future Research Directions

Nature-Inspired Algorithms for Identifying and Detecting Criminal Activities in the Internet of Things: A Review of Methods and Future Research Directions

Alireza Ghorbani Koundoghdi1 Nima Jafari Navimipor2

1) Department of Computer Engineering, Faculty of Engineering, Islamic Azad University, Tabriz Branch, Tabriz, Iran
2) Department of Computer Engineering, Faculty of Engineering, Islamic Azad University, Tabriz Branch, Tabriz, Iran

IOI Code : AAHD-ABHFCD
Publication : National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
    The rapid growth of the Internet of Things has connected millions of sensors and smart devices to the Internet. The resulting stream of sensor data can support urban security monitoring and the detection of suspicious events. This data, however, is large, heterogeneous, and high-dimensional. Real-time detection of criminal activity is therefore impractical unless a suitable subset of features is selected first, and feature selection is itself NP-hard because the search space grows exponentially with the number of features. Existing surveys have concentrated on intrusion detection and network security, or have treated metaheuristic algorithms only as companions to machine learning and deep learning methods. Reviews that examine the independent role of nature-inspired algorithms in crime detection from IoT data are rare. This paper addresses that gap by reviewing seven algorithms: the Genetic Algorithm, Particle Swarm Optimization, the Grey Wolf Optimizer, the Artificial Bee Colony, the Bat Algorithm, Cuckoo Search, and the Gorilla Troops Optimizer. For each algorithm, we describe its general mechanism, summarize recent studies, and discuss its advantages and limitations. The Genetic Algorithm, Particle Swarm Optimization, and the Grey Wolf Optimizer appear most often in recent work. In contrast, the Bat Algorithm and Cuckoo Search remain little tested despite their strong global search ability. No algorithm is superior in every case; the appropriate choice depends on the problem, from the type of data to the computational resources available. We close with suggestions for future research, including evaluation of these algorithms in real operational settings and the construction of dedicated crime detection datasets.

Keywords : internet of things nature inspired algorithms feature selection criminal activity detection metaheuristic algorithms