Article Per Year (5 Year)
p-Index From 2021 - 2026
0.444
P-Index
This Author published in this journals
All Journal Journal of Information Systems Engineering and Business Intelligence JOIV : International Journal on Informatics Visualization
Nguyen, Minh Duc
Unknown Affiliation
Computer Science & IT

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search
Journal : JOIV : International Journal on Informatics Visualization

 1 
Development of Intelligent Marine Logistics Models Using Machine Learning Nguyen, Minh Duc; Nguyen, P. Quy Phong; Luong Ha, Chuc Quynh; Dinh, Xuan Manh; Cao, Van Sam; Dat Do, Hoang
JOIV : International Journal on Informatics Visualization Vol 9, No 1 (2025)
Publisher : Society of Visual Informatics

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62527/joiv.9.1.3666

Abstract

This study looks into the development of intelligent maritime logistics models that use machine learning approaches to forecast crucial metrics like fuel usage and port delays. A comprehensive dataset assessed five machine learning models: Linear Regression, Decision Tree, Random Forest, XGBoost, and AdaBoost. Predictive capacities were assessed using key performance measures such as R², MSE, and MAPE. The results show significant heterogeneity in model performance, with Linear Regression attaining a modest test R² of 0.6845 for fuel prediction and 0.8831 for port delay prediction but suffering from high MSE (58745.23) and MAPE (26.90% for fuel). The Decision Tree showed significant overfitting, with a perfect R² (1.000) on training but decreasing to 0.7743 for fuel and 0.9880 for port delay on testing. Random Forest demonstrated balanced performance, with test R² values of 0.7598 for fuel and 0.9548 for port delay. MAPE values were also lower (23.66% for fuel and 5.66% for port delay). The best-performing model was XGBoost, with near-perfect test R² values of 0.7439 for fuel and 0.9880 for port delay, as well as a low MSE (39579.79 for fuel and 0.23 for port delay). AdaBoost produced comparable but somewhat lower results, with test R² values of 0.7188 for fuel and 0.9485 for port delay. These findings demonstrate XGBoost's strength in capturing nonlinear interactions and making solid predictions, whereas ensemble approaches outperform simpler models such as Linear Regression.
Co-Authors Cao, Van Sam Dat Do, Hoang Dinh, Xuan Manh Luong Ha, Chuc Quynh Nguyen, P. Quy Phong