MaTCH: An AI-powered application that allows for aggregating microplastics across studies

by

Editors' notes

This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:

fact-checked

peer-reviewed publication

trusted source

proofread

Change in morphological (a) and polymeric (b) categorical variables before and after semantic alignment. Illustration of hierarchical lumping via sunburst plot with 95% confidence intervals. (PP: polypropylene, PET: polyethylene terephthalate, PE: polyethylene, PA: polyamide, PS: polystyrene, PVC: polyvinyl chloride, and PPS: polyphenylene sulfide). Credit: Environmental Science & Technology (2024). DOI: 10.1021/acs.est.4c02406

A trio of environmental researchers at the University of California, Riverside, has developed an AI-based application that can be used for aggregating microplastics across studies. In their paper published in the journal Environmental Science & Technology, Hannah Hapich, Win Cowger and Andrew Gray describe how they applied machine learning to help classify microplastic bits to develop an application that researchers can use in a variety of fields to categorize microplastics in meaningful and useful ways.

In this new effort, the research trio noted that a lot of research has been done of late, looking into problems associated with microplastics—such studies have shown that the tiny particles have made their way into every conceivable part of the environment, including people's bodies, across the globe.

However, something else they noticed was that the way various researchers referred to the microplastics they were finding did not appear to follow any logical categorization system—no data standards were being used. Noting that something needed to be created to make sense of all the data, they created a system of their own—one based on artificial intelligence.

Microplastics in the environment are made up of different kinds of basic materials and are different in size, shape and color. Some are also known to cause certain ailments in certain animals. To create their categorization system, the researchers first had to find a way to convert existing data from information describing size, shape and material, to mass, volume and density.

Abstract. Credit: Environmental Science & Technology (2024). DOI: 10.1021/acs.est.4c02406

Doing so allowed for what the team describes as "harmonizing" the data, which makes the study and comparison of seemingly incompatible datasets possible. They then used a machine learning app to find the different words used by researchers in their papers to describe the microplastics they were finding and then to convert them to standard terminology and measurements. They have named the completed tool Microplastics and Trash Cleaning and Harmonization (MaTCH).

In addition to categorization, the researchers also added useful output features, such as sunburst plots that show data in top-down fashion, and how different types of microplastics found in a given research effort can be compared with those found by researchers working on other efforts.

More information: Hannah Hapich et al, Microplastics and Trash Cleaning and Harmonization (MaTCH): Semantic Data Ingestion and Harmonization Using Artificial Intelligence, Environmental Science & Technology (2024). DOI: 10.1021/acs.est.4c02406

Journal information: Environmental Science & Technology

© 2024 Science X Network