Multi-Label Classification for AIS Data Anomaly Detection Using Wavelet Transform

Thanks to the Automatic Identification System (AIS), ships and other maritime equipment are able communicate with each other, for example, by sending information about their position. This solution allows early collision detection when two or more on a course. In newer version of AIS, satellite infrastructure is used extend communication range. Unfortunately, AIS deals so-called packet effect: since there problem synchronizing data coming from multiple terrestrial areas, single may receive several messages at same time be unable correctly process them, causing get lost garbled. this article, machine learning based framework detecting incorrect presented. approach, after first stage (clustering), dedicated anomaly algorithm searches damaged conducts multi-label classification (with Random Forest wavelet transform) decide which fields such message requires further correction. The results measuring effectiveness proposed approach using real