Machine learning in analytical chemistry for cultural heritage: A comprehensive review

Abstract

In recent years, machine learning (ML) has gained significant importance in the field of cultural heritage research. Its advanced data analysis techniques have become a crucial tool in many areas of heritage science. This literature review intends to discuss the applications of ML to studies on cultural heritage objects using the analytical chemistry methods. The analysis of large datasets obtained from complex measurements with the use of ML algorithms has been demonstrated to result in a deeper understanding of the studied objects. Such analyses have also been shown to provide new perspectives on many problems. The article outlines studies on varied materials such as pigments, paper, metals, and ceramics. It presents analyses that use diverse ML methods, including unsupervised and supervised techniques, utilizing both traditional algorithms and neural networks. It also provides an introduction to understanding ML, its principles and methods, with the focus on practices applicable to heritage science.