Knowledge discovery using genetic algorithm for maritime situational awareness

Due to the large volume of data related to vessels, to manually pore through and to analyze the information in a bid to identify potential maritime threat is tedious, if at all possible. This study aims to enhance maritime situational awareness through the use of computational intelligence techniques in detecting anomalies. A knowledge discovery system based on genetic algorithm termed as GeMASS was proposed and investigated in this research. In the development of GeMASS, a machine learning approach was applied to discover knowledge that is applicable in characterizing maritime security threats. Such knowledge is often implicit in datasets and difficult to discover by human analysts. As the knowledge relevant to maritime security may vary from time to time, GeMASS was specified to learn from streaming data and to generate up-to-date knowledge in a dynamic fashion. Based on the knowledge discovered, the system functions to screen vessels for anomalies in real-time. Traditionally in maritime security studies, datasets that are applied as knowledge sources are related to vessels’ geographical and movement information. This study investigated a novel leverage of multiple data sources, including Automatic Identification System, classification societies, and port management and security systems for the enhancement of maritime security. A prototype of GeMASS was developed and employed as a vehicle to study and demonstrate the functions of the proposed methodology.