Abstract
This study deals with short-term weather forecasting in maritime environments using machine learning. Data from autonomous ship sensors, such as, temperature, humidity, wind speed, electrostatic field, and sea conditions, were used to enhance forecast accuracy. A Radial Basis Function (RBF) model was developed to tackle nonlinear structure of the data. Data preprocessing and model development were performed using MATLAB, while model evaluation was conducted in Python. The dynamic modeling approach that was adopted demonstrated significant accuracy improvements over static models, indicating the potential for enhancing maritime safety and operational efficiency with automated, energy-efficient systems.
