Selective mechanochemical conversion of post-consumer polyethylene terephthalate waste into hcp and fcu UiO-66 metal–organic frameworks

Summary

Machine learning models including SVM, k-NN, random forest, and logistic regression were evaluated for classifying ABS, polypropylene, and Nylon-66 polymers using FTIR spectral data, with k-NN achieving the best overall accuracy despite challenges distinguishing spectrally similar materials. Accurate polymer identification via FTIR and ML is foundational to microplastic research, where rapid automated classification of plastic types in environmental samples is a persistent analytical bottleneck.