Premium
Prediction models of rice cooking quality
Author(s) -
Borries George,
Bassinello Priscila Z.,
Rios Érica S.,
Koakuzu Selma N.,
Carvalho Rosangela N.
Publication year - 2018
Publication title -
cereal chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.558
H-Index - 100
eISSN - 1943-3638
pISSN - 0009-0352
DOI - 10.1002/cche.10017
Subject(s) - principal component analysis , sensory system , novelty , texture (cosmology) , logistic regression , pattern recognition (psychology) , statistics , principal component regression , artificial intelligence , linear regression , biological system , mathematics , chemistry , computer science , psychology , cognitive psychology , social psychology , image (mathematics) , biology
Background and objectives Rice quality can be primarily assessed by evaluating its texture after cooking. The classical sensory evaluation is an expensive and time‐consuming method as it requires training, capability, and availability of people. Therefore, this study investigated the possibility of replacing sensory evaluation by analyzing the relationship between sensory and instrumental texture and viscosity measurements. Findings Models predicting the sensory evaluation were developed by applying statistical methods such as principal component analysis and polytomous logistic regression. The level of prediction efficiency of these models was obtained by estimating the apparent misclassification error rate and also using the ROC curve graph. The results indicated that the instrumental texture measurements were consistently related to sensory analysis. Similarly, viscosity measurements enabled the prediction of results obtained by sensory texture evaluation. Conclusions Principal component analysis together with polytomous logistic regression is an efficient method to predict sensorial stickiness of rice using viscosity measures of texture as predictors. Significance and novelty The current study was able to correctly predict sensory stickiness in about 86% of cases using just one principal component formed by a combination of measures of apparent amylose content, gelatinization temperature, and RVA parameters.