Using NMR spectroscopy and linear discriminant analysis to molecular profile varietal honey

<p>In recent years, varietal honey has been a massive target of adulteration through mislabeling and the addition of other sugars. Unethical companies do this to cut production costs while still charging the consumer full price. Previous studies have used nuclear magnetic resonance (NMR) to unravel possible adulteration in honey, but currently, there are no standard rapid methods to authenticate varietal honey. In our research, we collected NMR signatures (“fingerprints”) and combined them with linear discriminant analysis (LDA) to predict the varietal, country, and region of varietal honey. As a part of the project, we investigated whether an adjustment to the spectral data based on a signal-to-noise cutoff would provide better predictive ability. Our results are inconclusive as to whether a signal-to-noise adjustment provides a meaningful advantage; however, they demonstrate that combining NMR data with LDA enables prediction of honey varietal, country, and region at rates significantly better than random chance.</p>

Name thesis:degree_name

Master of Science (MS)

Level thesis:degree_level

Open Access Thesis

Discipline thesis:degree_discipline

Chemistry

Year dc:date.available

2026

Author dc:creator

• Mac, Taylor

Contributors dc:contributor

• Cory Emal, Ph.D.
• Heather Holmes, Ph.D.
• Gregg Wilmes, Ph.D.

• Chemistry
• Organic Chemistry

Repository record dc:identifier

https://commons.emich.edu/theses/1335

OAI identifier oai:identifier

oai:commons.emich.edu:theses-2680