Coating matching recommendation based on improved fuzzy comprehensive evaluation and collaborative filtering...

Area of Science:

• Marine Engineering
• Materials Science
• Computational Intelligence

Background:

• Ship coating selection is complex, influenced by corrosive environments, durability needs, and specific application areas.
• Current reliance on expert experience for coating matching hinders scientific management and cost control in shipbuilding.

Purpose of the Study:

• To develop an intelligent, data-driven approach for optimizing ship coating matching.
• To enhance the scientific basis and cost-effectiveness of ship painting processes.

Main Methods:

• A hybrid algorithm (IFCE-CF) combining fuzzy comprehensive evaluation and improved collaborative filtering.
• Analytic hierarchy process (AHP) to establish a coating matching evaluation index system.
• User label weighting to optimize fuzzy comprehensive evaluation and matrix decomposition for collaborative filtering.

Main Results:

• The proposed IFCE-CF algorithm achieved a root mean square error (RMSE) of < 1.02 and a mean absolute error (MAE) of < 0.75.
• Demonstrated significantly improved prediction accuracy compared to existing research methods.
• Validated using historical shipyard coating data from 2010-2020.

Conclusions:

• The IFCE-CF algorithm effectively provides accurate coating matching recommendations.
• This intelligent approach offers a more scientific and cost-efficient solution for ship coating design.
• The method proves valuable for improving shipbuilding process management and cost control.