master's thesis
Chemometric dependence model of quality of wheat grain and flour

Iva Vasilj (2016)
Josip Juraj Strossmayer University of Osijek
FACULTY OF FOOD TECHNOLOGY
Department of Process Engineering
Sub-department of Modelling, Optimisation and Automatisation
Metadata
TitleKemometrijski model zavisnosti svojstava kvalitete pšeničnog zrna i brašna
AuthorIva Vasilj
Mentor(s)Damir Magdić (thesis advisor)
Daniela Horvat (thesis advisor)
Abstract
Kemometrijskim metodama su određene povezanosti između svojstava pšeničnog zrna i brašna od 24 kultivara. Kultivari su uzgajani na poljima Poljoprivrednog instituta Osijek tijekom deset godina (2005.-2014.). Analizirana svojstva su: prinos pšenice (Y), hektolitarska masa (HL), masa tisuću zrna pšenice (TKW), udio proteina (P), udio vlažnog glutena (WG), gluten indeks (GI), sedimentacijska vrijednost (SED), broj padanja (FN) i izbrašnjavanje (FY). Na izmjerenom setu podataka provedena je sljedeća deskriptivna statistička analiza: određivanje srednjih vrijednosti, medijana, minimalne i maksimalne vrijednosti, standardne devijacije, koeficijenta varijabilnosti i njihovih međusobnih korelacijskih koeficijenata. Provedene su i sljedeće kemometrijske analize: analiza glavnih komponenti (PCA), klaster analiza (CA) i regresijska analiza metodom najmanjih kvadrata (PLSR). Analiza glavnih komponenti korištena je za smanjenje broja varijabli, klaster analiza korištena je za pokazivanje veza među svojstvima, a regresijska analiza metodom najmanjih kvadrata korištena je za izradu prediktivnih modela. Rezultati pokazuju da je na temelju snažnih korelacija moguće smanjiti broj varijabli i sa manje od devet svojstava opisati varijabilnost seta podataka. Predloženim prediktivnim matematičkim modelima moguće je korištenjem izmjerenih vrijednosti izračunavati vrijednosti preostalih svojstava s točnošću većom od 90%.
Keywordswheat flour chemometric methods
Parallel title (English)Chemometric dependence model of quality of wheat grain and flour
Committee MembersSandra Budžaki (committee chairperson)
Damir Magdić (committee member)
Daniela Horvat (committee member)
Ana Bucić-Kojić (committee member)
GranterJosip Juraj Strossmayer University of Osijek
FACULTY OF FOOD TECHNOLOGY
Lower level organizational unitsDepartment of Process Engineering
Sub-department of Modelling, Optimisation and Automatisation
PlaceOsijek
StateCroatia
Scientific field, discipline, subdisciplineBIOTECHNICAL SCIENCES
Food Technology
Engineering
Study programme typeuniversity
Study levelgraduate
Study programmeProcess Engineering
Academic title abbreviationmag. ing. proc.
Genremaster's thesis
Language Croatian
Defense date2016-10-10
Parallel abstract (English)
Chemometric methods were used to determine connection among properties of wheat grain and flour produced from 24 cultivars. Cultivars were breed on Agricultural institute Osijek in period of ten years (2005-2014). Following properties were analysed: yield (Y), hectolitre mass (HL), thousand kernel weight (TKW), protein content (P), wet gluten content (WG), gluten index value (GI), sedimentation value (SED), falling number (FN) and flour yield (FY). Following descriptive statistical analysis was applied on experimental set of data: mean value, median, minimum and maximum, standard deviation, coefficient of variability and coefficient of correlation. Following chemometric analysis were applied: principal component analysis (PCA), cluster analysis (CA) and partial least square method (PLSR). Principal component analysis was used for reduction of number of variables, cluster analysis was used for enlightening connections among properties while partial least square regression analysis was used for designing predictive mathematical models. Results show that strong correlations can be used for decreasing number of variables and describing variability of data set with less than nine properties. Using proposed predictive mathematical models based on experimental data makes possible calculating values of rest of properties with precision bigger than 90%.
Parallel keywords (Croatian)pšenica brašno kemometrijske metode
Resource typetext
Access conditionOpen access
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:109:501144
CommitterIvana Šuvak