Primjena kemometrije i umjetnuih neuronskih mreža u karakterizaciji in- vitro procesa probave multivitaminskih dodataka prehrani

master's thesis

• Title (english): Applications of chemometrics and artificial neural networks in characterization in- vitro digestion process of multivitamin dietary supplements
• Author: Josip Morduš
• Mentor: Davor Valinger (mentor)
• 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: 2019-09-25, Croatia
• Scientific / art field, discipline and subdiscipline: BIOTECHNICAL SCIENCES; Nutrition

Abstract

Multivitaminski dodaci prehrani predstavljaju jedan od najkonzumiranijih oblika tih proizvoda i na tržištu se mogu naći u brojnim izvedbama. Jedan od takvih oblika je i multivitaminski prah koji se konzumira kao napitak nakon miješanja s vodom. Velik je izazov provesti suvremene računalne analize probavljivosti u kontekstu antioksidativne aktivnosti te brojnih drugih svojstava u svrhu karakterizacije potencijalnog zdravstvenog učinka, a upravo takva metoda je primjenjivana u ovom radu. Cilj ovoga rada bio je, uz određivanje fizikalnih (raspodjela veličine čestica, nasipna gustoća, svojstva tečenja, sadržaj vlage, električna vodljivost i ukupne otopljene tvari) i kemijskih (antioksidacijski kapacitet, sadržaj šećera) svojstava multivitaminskih dodataka prehrani, tijekom in-vitro probave napraviti modele umjetnih neuronskih mreža (ANN), na temelju prethodno napravljene analize glavnih komponenata (PCA), koji bi povezali uzorke snimljene blisko infracrvenom spektroskopijom (NIR) i ultraljubičasto-vidljivom (UV/VIS) spektroskopijom s dobivenim rezultatima. Kreirani ANN model in-vitro procesa probave puno se bolje pokazao u slučaju NIR spektroskopije gdje su vrijednosti koeficijenata determinacije za predviđanja vrijednosti električne vodljivosti, refrakcijskog indeksa, stupnja Brix-a te antioksidativne aktivnosti bile u rasponu 0,930 – 0,997 dok su za UV/VIS spektroskopiju bile u rasponu 0,115 -0,449.

Abstract (english)

Multivitamin dietary supplements represent one of the most consumed types of those products, which can be found on the market in numerous forms. One of these forms is a multivitamin powder which is consumed as a beverage after mixing it with water. It is challenging to perform modern computational analyses of digestibility in the context of antioxidant activity and many other properties to characterize the potential health effect, and in this thesis, that exact method is used. The aim of this study was to determine the physical (particle size distribution, bulk density, flow properties, moisture content, electrical conductivity, and total dissolved solids) and chemical (antioxidant capacity, sugar content) properties of multivitamin supplements, additionally to create artificial neural network (ANN) models during in-vitro digestion, based on previously performed principal component analysis (PCA), which would connect samples recorded by near-infrared (NIR) spectroscopy and ultraviolet-visible (UV/VIS) spectroscopy with the obtained results. Created ANN model of the in-vitro digestion process proved much better in the case of NIR spectroscopy where the values of the coefficients of determination for the predictions of the values of electrical conductivity, refractive index, degree of Brix and antioxidant activity were ranged from 0,930 to 0,997 while for UV/VIS spectroscopy were ranged from 0,115 to 0,449.

• Keywords: multivitamin; dodatak prehrani; umjetne neuronske mreže (ANN); in-vitro probava
• Keywords (english): multivitamin; dietary supplement; artificial neural network (ANN); in-vitro digestion
• Language: croatian
• URN:NBN: urn:nbn:hr:159:348685
• 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: 2020-01-28 14:41:22