TY - JOUR
T1 - Environmental sustainability
T2 - A machine learning approach for cost analysis in plastic recycling classification
AU - Carrera, Berny
AU - Mata, Judit Bazin
AU - Piñol, Victor Luid
AU - Kim, Kwanho
N1 - Publisher Copyright:
© 2023 Elsevier B.V.
PY - 2023/10
Y1 - 2023/10
N2 - Recycling plastics can reduce waste generation and improve waste management, but the recycling industry needs both cost reduction and increased revenue to be economically viable. Recently, recycling plastic classification techniques with Artificial Intelligence have gained popularity, as they can avoid manual sorting, which is time-consuming and economically less profitable than automatidc processing. In this paper, we provide an economic framework for quality sorting control by classifying plastics based on the infrared spectrum of polymers and machine learning algorithms. In addition, the suggested framework offers a method for selecting the algorithm according to the polymer's income class and the highest economic advantages. Furthermore, our experiments probe that Fourier-transform infrared (FTIR) and near-infrared (NIR) spectroscopies combined with machine learning algorithms are suitable for plastic classification as four datasets and seven machine learning algorithms have been tested to classify Polyethylene (PE), Polypropylene (PP), Polyethylene terephthalate (PET), polystyrene (PS), and Polyvinyl chloride (PVC).
AB - Recycling plastics can reduce waste generation and improve waste management, but the recycling industry needs both cost reduction and increased revenue to be economically viable. Recently, recycling plastic classification techniques with Artificial Intelligence have gained popularity, as they can avoid manual sorting, which is time-consuming and economically less profitable than automatidc processing. In this paper, we provide an economic framework for quality sorting control by classifying plastics based on the infrared spectrum of polymers and machine learning algorithms. In addition, the suggested framework offers a method for selecting the algorithm according to the polymer's income class and the highest economic advantages. Furthermore, our experiments probe that Fourier-transform infrared (FTIR) and near-infrared (NIR) spectroscopies combined with machine learning algorithms are suitable for plastic classification as four datasets and seven machine learning algorithms have been tested to classify Polyethylene (PE), Polypropylene (PP), Polyethylene terephthalate (PET), polystyrene (PS), and Polyvinyl chloride (PVC).
KW - Environment
KW - Machine learning classification
KW - Optimization
KW - Plastic recycling industry
KW - Plastic recycling revenue
KW - Profitable plastics
KW - Quality sorting
KW - Sustainable cities
KW - Waste management
UR - http://www.scopus.com/inward/record.url?scp=85162986322&partnerID=8YFLogxK
U2 - 10.1016/j.resconrec.2023.107095
DO - 10.1016/j.resconrec.2023.107095
M3 - Article
AN - SCOPUS:85162986322
SN - 0921-3449
VL - 197
JO - Resources, Conservation and Recycling
JF - Resources, Conservation and Recycling
M1 - 107095
ER -