Every year, the University of Ljubljana selects the ten most prominent research achievements among the proposals received from several members of the University, based on aspects, including impact, comprehensiveness, and applicability. This year, the research into the dating of historical books using near-infrared (NIR) spectroscopic data and machine learning was chosen as one of the ten best research achievements.
For the first time, Dr Floriana Coppola, post-doctoral researcher at the Heritage Science Lab Ljubljana, and her co-authors systematically explored the uncertainty associated with the use of NIR spectroscopic data combined with machine learning to predict the publication dates of paper books dated between 1851 and 2000. The study was published in the eminent Journal of the American Chemical Society, and was also summarised as research highlight in the prestigious Nature Reviews Chemistry. The research involved the acquisition of 3000 spectra of the books from the collections of the National and University Library of Slovenia, and analysis using three machine learning algorithms. The results show that paper can be dated extremely precisely (within two years), and compositional changes are more important for accuracy of prediction than paper degradation. Such breakthrough results were achieved thanks to the intensive and interdisciplinary collaboration between researchers from the Heritage Science Lab Ljubljana (Dr Floriana Coppola, Dr Jernej Markelj, Prof Matija Strlič), the National and University Library of Slovenia (Dr Jasna Malešič) and the Department of Economics and Business Sciences of the University of Cagliari, Italy (Dr Luca Frigau, Prof Dr Claudio Conversano).
Funding: This research is part of the UNCERTIR project that has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska Curie grant agreement No. 101032212. Additional support by the Slovenian Research Agency: projects I0-E012, J4-3085, N1-0271, and P1-0153.