Marine accident learning with fuzzy cognitive maps (MALFCMs)

Area of Science:

• Maritime Safety
• Accident Analysis
• Risk Management

Background:

• Statistical analysis of maritime accidents reveals trends but lacks methods to model complex factor interrelations.
• Existing accident databases offer insufficient detail to determine the importance of individual contributing factors.
• This gap hinders accurate risk assessments and the prediction of risk control effectiveness.

Purpose of the Study:

• To introduce a novel technique, Marine Accident Learning with Fuzzy Cognitive Maps (MALFCMs), for modeling maritime accident contributors.
• To develop a method capable of estimating the importance of each contributing factor in maritime accidents.
• To enhance the objectivity and accuracy of risk assessments by providing a robust factor weighting technique.

Main Methods:

• Application of Fuzzy Cognitive Maps (FCMs) to model relationships between maritime accident contributors.
• Direct learning from maritime accident databases combined with the integration of expert opinions.
• Development of the MALFCM method to overcome the inherent subjectivity of traditional FCMs.

Main Results:

• A novel MALFCM method has been successfully developed to weight human-contributing factors in maritime accidents.
• The MALFCM method overcomes the primary disadvantage of traditional FCMs by reducing subjectivity.
• The method demonstrates the ability to produce logical results even when relying solely on historical data without expert input.

Conclusions:

• MALFCM offers a more objective and reliable approach to understanding and quantifying factors in maritime accidents.
• This technique can significantly improve the precision of maritime risk assessments and the evaluation of safety interventions.
• MALFCM provides a valuable tool for maritime safety research, capable of utilizing diverse data sources effectively.