<![CDATA[In this study, the author presents the findings of a systematic literature review that includes articles published in the last ten years in which the author explains the application of Machine Learning (ML) techniques and methods to sensing technology in support of maritime defense. By searching in the Scopus and Mendeley databases, the authors took, filtered, and analyzed the contents of 30 articles, and encoded these resources following repeated methods in the Grounded Theory approach. According to the literature review, the most widely used ML model is the CNN model with 16 article titles and has the best performance quality level of 99.79%. This is because CNN is highly effective at classification. After all, it can identify intricate patterns and relationships between features. However, from the perspective of maritime defense technology, which focuses on threats from underwater, an article that analyzes ship noise for ten articles and submarine noise for two articles takes as its subject. This is caused by both data confidentiality and data limitations. To solve the problem, previous researchers used the data pre-processing method of Image Data Augmentation. However, despite improving the performance of the ML model, it has the potential for a huge over-fitting phenomenon.]]>
Copyrights © 2024
