Exploring Maritime Movement Information: An Explainable AI Approach using Hi-DBSCAN and SHAP

Authors

  • Nitin Newaliya
  • Vikas Siwach
  • Harkesh Sehrawat
  • Yudhvir Singh

DOI:

https://doi.org/10.6977/IJoSI.202412_8(4).0009

Keywords:

Data Analytics, DBSCAN, Maritime Situational Awareness, Explainable AI, AIS, SHAP

Abstract

Maritime movement information is pivotal for several applications, including monitoring and examining vessel activities, ensuring efficient and secure navigation, logistics optimisation, and enhancing safety and environmental protection. The maritime industry relies on Automatic Identification System (AIS) data, which provides information on movement of vessels at sea. Determining meaningful and useful insights from this data is a challenge. The complexity and volume of the information make it difficult for traditional methods to provide in-depth insights and explanations. This paper presents an innovative Explainable AI (XAI) approach to explore maritime movement information using AIS data by leveraging high dimensional Density-Based Spatial Clustering of Applications with Noise (Hi-DBSCAN) algorithm and SHAP (SHapley Additive exPlanations) values in a novel way. The effectiveness of this approach in extracting meaningful insights from maritime movement information is demonstrated experimentally, which also provides transparency and explainability, empowering maritime stakeholders to gain a deeper understanding and make data-driven decisions.

Downloads

Published

2024-12-30