Revealing factors influencing polymer degradation with rank-based machine learning

Area of Science:

• Environmental Science
• Materials Science
• Data Science

Background:

• Marine sustainability is threatened by inefficient polymer waste treatment.
• Understanding polymer degradability is crucial for designing sustainable materials.
• Limited and diverse degradability datasets hinder comprehensive analysis.

Purpose of the Study:

• To develop a machine learning platform for polymer degradability evaluation.
• To integrate heterogeneous datasets for large-scale degradability analysis.
• To identify key factors influencing polymer degradability.

Main Methods:

• Utilized a rank-based machine learning technique (RankSVM).
• Developed a ranking model by integrating three diverse polymer degradability datasets.
• Employed decision tree analysis on the ranking model.

Main Results:

• Successfully created a platform for evaluating polymer degradability.
• Established a robust ranking model for polymer degradability.
• Identified dominant factors controlling polymer degradability through analysis.

Conclusions:

• The developed platform enables effective analysis of diverse polymer degradability data.
• The identified factors can guide the design of more degradable polymer materials.
• This approach contributes to addressing the challenge of polymer waste in marine environments.