Breakthrough: NIR and Machine Learning for Dating of Historical Books

Heritage scientists can now accurately date historical paper using just infrared light and complex computational analysis.

In a piece of research published in the top-tier Journal of the American Chemical Society, the post-doctoral researcher dr. Floriana Coppola and colleagues have shown that exceptional dating accuracy is possible of only 2 years – this is not achievable with any other method and certainly not with any other non-destructive method.

Fig. 1: Dr. Floriana Coppola during NIR spectroscopic analyses.


Floriana analysed 100 books from the general collection of the National and University Library of Slovenia, dated from 1851-2000, an exceptional transitional period in the history of papermaking. A total of 3000 NIR spectra were measured and combined with three machine learning algorithms to understand whether and how paper variability and natural degradation influence the dating models. It was shown that the common spectral features of cellulose and protein structures, are of importance for the predictions rather than degradation that does not meaningfully influence the prediction accuracy

Fig. 2: Schematic representation of the study, exploring the interpretative and predictive ability of dating models combining NIR spectroscopic data and machine learning.


The research was a collaboration between researchers from the Heritage Science Lab Ljubljana (Floriana Coppola, Jernej Markelj, Matija Strlič), the Department of Business and Economics of the University of Cagliari, Italy (Luca Frigau, Claudio Conversano) and the National and University Library of Slovenia (Jasna Malešič).

Read more about the research in the journal article Near-Infrared Spectroscopy and Machine Learning for Accurate Dating of Historical Books”.



The work was conducted within the UNCERTIR project (European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska Curie grant agreement No 101032212). We acknowledge additional support by the Slovenian Research Agency, projects J4-3085, N1-0271, and P1-0153.

Please contact Floriana for further information.

Cover_lezeci-min obr
Fig. 3: Cover design by Nina Cavz.

Leave a Comment

Your email address will not be published. Required fields are marked *