Random Forest Analysis of X-ray Diffraction and Scattering Data on Crystalline Polymer

Crystalline polymers have a hierarchical structure in which polymer chains are folded. Although each hierarchical structure strongly affects the physical properties of crystalline polymers, it is hard to describe the relationship between the formation conditions, crystal structure and physical properties. We used Random Forest regression to comprehensively investigate the relationship between these features of polylactic acid (PLA), a biodegradable crystalline polymer. It was suggested that important features for mechanical property and biodegradability, where the trade-off relationship between them is a significant issue of PLA, are related to the different level crystal structures. This shows that it is possible to use Random Forest for complex prediction of crystalline polymer properties to search for important forming conditions and crystal structures.