Machine Learning-Based Data Quality Assessment for the Textile and Clothing Digital Product Passport

Transparency in business practices is essential for sustainability, ensuring that resources are used responsibly and that environmental and social impacts are properly measured and monitored, allowing the end consumer to make informed purchasing decisions without feeling cheated. The Digital Product Passport (DPP) promotes transparency by providing detailed information about a product’s origin, composition, and life-cycle activities, enabling more sustainable and responsible choices. The implementation of the DPP for textile and clothing items faces many challenges due to the large number and diversity of companies involved in the value chain of these products, combined with the large amount and variability of information that needs to be collected. Therefore, the integration and standardization of data from these companies is one of the largest present challenges. In this article, we study the use of Machine Learning (ML) algorithms for validating, in a homogeneous way, the quality of the data submitted by each company for the implementation of the DPP. We have studied four solutions that, using datasets organized in different ways and using different ML algorithms, enable selecting the solution that best suits each particular situation.