본문 바로가기
  • Home

Analysis of Data Design Factors Affecting Ship Fuel Consumption Prediction: A Multi-Factor Experimental Approach

  • Journal of Internet of Things and Convergence
  • Abbr : JKIOTS
  • 2025, 11(6), 23
  • Publisher : The Korea Internet of Things Society
  • Research Area : Engineering > Computer Science > Internet Information Processing
  • Received : December 2, 2025
  • Accepted : December 22, 2025
  • Published : December 31, 2025

Bong-Min Cha 1 Heung-Jun Im 2 Sang-Bong LEE 3 Seongpil Kang 3 Dong-Hyun Kim 4

1국립부경대학교
2국립부경대학교 산업및데이터공학과
3랩오투원
4국립부경대학교 미래융합학부

Accredited

ABSTRACT

With the strengthening global commitment to a sustainable future, efforts to achieve net-zero greenhouse gas emissions by 2050 are accelerating. Accordingly, the need for decarbonization in the maritime sector, which carries over 90% of global trade and produces about 3% of anthropogenic emissions, continues to grow. Yet many existing studies rely on post-voyage variables or assume temporal continuity and data completeness, limiting their applicability to real operations. This study proposes an automated experimental framework that predicts fuel oil consumption using only variables available at the voyage planning stage. Key experimental factors—normal-operation filtering, input representation, temporal resolution, data-splitting methods, target variable selection, and model type— are systematically defined. The Hydra framework automates experiments across factor combinations, and ANOVA quantifies each factor’s contribution to prediction performance. Results show that operational condition settings strongly influence accuracy, and the study provides a data design strategy to support future energy-efficient voyage planning.

Citation status

* References for papers published after 2024 are currently being built.