Premium
Predicting oxidative stability of vegetable oils using neural network system and endogenous oil components
Author(s) -
Przybylski Roman,
Zambiazi Rui C.
Publication year - 2000
Publication title -
journal of the american oil chemists' society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.512
H-Index - 117
eISSN - 1558-9331
pISSN - 0003-021X
DOI - 10.1007/s11746-000-0146-x
Subject(s) - vegetable oil , predictability , composition (language) , food science , chemistry , tocopherol , artificial neural network , chemical composition , antioxidant , biochemistry , organic chemistry , vitamin e , mathematics , computer science , artificial intelligence , linguistics , statistics , philosophy
The usefulness of Artificial Neural Network Systems (ANNW) to predict the stability of vegetable oil based on chemical composition was evaluated. The training set, comprised of a composition of major and minor components of vegetable oil as inputs and as outputs, induction period and values of slopes for initiation and propagation, was measured by oxygen consumption. The best predictability was achieved for oils stored at 35°C with light exposure, when the major fatty acids, chlorophylls, tocopherols, tocotrienols, and metals were used as predictors. For oils stored at 65°C without light, a good predictability was obtained when composition of the major fatty acids and the amounts of tocopherols and tocotrienols were used. These results suggest that vegetable oil stability can be successfully predicted by ANNW when partial oil composition is known.