Prediction of Ship Time in Port Using Machine Learning Algorithm

Abstract

Effective time management is crucial in the shipping business at ports, ensuring smooth operations from a ship’s arrival to its departure. Delays in port can disrupt sailing schedules, leading to inefficiencies in logistics and increased operational costs. This study aims to predict the duration of a ship's stay in port, focusing on container and general cargo ships. Accurate predictions can help optimize scheduling and resource allocation. The research applies machine learning techniques, utilizing historical arrival and departure data from Tanjung Priok Port for 2022 and 2023. Four algorithms were evaluated: Random Forest, Linear Regression, K-Nearest Neighbors (KNN) Regression, and Support Vector Machine (SVM). Results indicate that Linear Regression provides the most accurate predictions, outperforming other models in terms of Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). Additionally, its coefficient of determination (R²) is closest to one, confirming its reliability for forecasting ship turnaround times.