A machine learning algorithm for high throughput identification of FTIR spectra: Application on microplastics collected in the Mediterranean Sea

This article proposes a protocol for analyzing the huge number of microplastics present in samples taken from the marine environment. This analysis currently requires a great deal of time and advanced skills. This new protocol therefore proposes an automated machine learning approach based on spectrophotometric analysis. First, the spectra of pure plastics are entered into the algorithm, which uses them to determine the chemical nature of the microplastics present in the sample.

This approach has proven to be highly effective for certain polymers. For other polymers, bias can be reduced by adding new reference spectra for these plastics to the algorithm’s database. Other studies have used a similar process to identify microplastics, so other spectrum databases are available. By sharing this data, the robustness and reliability of the algorithm could be quickly improved. Nevertheless, it has already enabled the analysis of more than 4,000 samples, confirming its effectiveness.

---