Detekcija patvorenja meda agavinim i javorovim sirupom primjenom blisko infracrvene spektroskopije

master's thesis

• Title (english): Detection of honey adulteration with agave and maple syrup using near infrared spectroscopy
• Author: Matea Stulić
• Mentor: Tamara Jurina (mentor)
• Committee member: Ana Jurinjak Tušek (predsjednik povjerenstva)
• Committee member: Maja Benković (član povjerenstva)
• Committee member: Irena Keser (član povjerenstva)
• Granter: University of Zagreb; Faculty of Food Technology and Biotechnology; (Department of Process Engineering); (Laboratory for Measurement, Regulation and Automatisation); Zagreb
• Defense date and country: 2024-09-26, Croatia
• Scientific / art field, discipline and subdiscipline: BIOTECHNICAL SCIENCES; Nutrition

Abstract

U okviru ovog rada analizirana su fizikalna svojstva 15 uzoraka meda bagrema s različitim udjelima komponenata za patvorenje (agava ili javor). Primijenjena je blisko infracrvena spektroskopija, u kombinaciji s kemometrijskim metodama, za razvijanje modela u svrhu predviđanja udjela dodane patvorine. Rezultati fizikalnih svojstava pokazuju da su uzorci izvornih medova u skladu s Pravilnikom o medu, osim udjela vode, a dodatak patvorine lako se može detektirati mjerenjem električne vodljivosti. Na temelju izračunatih vrijednosti koeficijenata determinacije i RMSE vrijednosti za predikciju, validaciju i kalibraciju može se zaključiti kako svi razvijeni modeli parcijalne regresije metodom najmanjih kvadrata nisu prikladni za primjenu na uzorcima koji nisu uključeni u model. Razvijene umjetne neuronske mreže osiguravaju dobro slaganje eksperimentalnih i modelom predviđenih vrijednosti na razini učenja, testiranja i validacije uz R2 ≥ 0,9435 i niske pogreške modela RMSE ≤ 0,0057, za patvorine s javorom, stoga se mogu primijeniti za predviđanje udjela dodane patvorine na temelju snimljenih NIR spektara.

Abstract (english)

The physical properties of 15 samples of acacia honey with different amounts of adulterant (agave or maple) were analysed. Near infrared spectroscopy, in combination with chemometrics, was applied for developing models to predict the amount of added adulterant. The results of physical properties showed that the original honey samples comply with the Regulations, except for the water content, and the addition of adulterants can be detected by measuring electrical conductivity. Based on the determination coefficients and RMSE for prediction, validation and calibration, it can be concluded that all developed partial least squares models are not suitable for samples not included in the model. The developed artificial neural networks provided good agreement between experimental and model predicted values at the level of learning, testing and validation, with R2 ≥ 0,9435 and RMSE ≤ 0,0057 for adulterants with maple, therefore they can be applied to predict the amount of added adulterant based on the recorded NIR spectra.

• Keywords: blisko infracrvena spektroskopija; kemometrija; patvorenje; med; umjetne neuronske mreže
• Keywords (english): near infrared spectroscopy; chemometrics; adulteration; honey; artificial neural networks
• Language: croatian
• URN:NBN: urn:nbn:hr:159:634319
• Study programme: Title: Nutrition; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra nutricionizma (magistar/magistra nutricionizma)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-10-30 13:30:21