AIS Data Pre-Processing for Trajectory Clustering Data Preparation

Automatic Identification System (AIS) is radio navigation equipment for a vessel that has been required by the International Maritime Organization (IMO). The AIS dataset contains vessel information and vessel position. Various analyses can utilize the availability of AIS's extensive data history. In those analyses, it is necessary to know the vessel's trajectory pattern. With the development of data mining techniques, vessel trajectory patterns can be obtained by clustering. However, AIS data cannot be directly used in the clustering process. Data pre-processing is required due to the complexity of the trajectory data and the need to reduce noises in AIS data with large sizes. This study proposes a pre-processing model with data cleaning, trajectory extraction, and trajectory compression stages. Results show that the proposed model can reduce noise and, at the same time, reduce rows that will affect the following clustering process.