Portal de Periódicos da CAPES

Artigo Acesso aberto Produção Nacional Revisado por pares

BIBTEX RIS

Use of multivariate statistical methods for classification of olive oil

2021; UNIVERSIDADE FEDERAL RURAL DE PERNAMBUCO; Volume: 6; Issue: 1 Linguagem: Inglês

10.24221/jeap.6.1.2021.2815.035-043

2525-815X

Moacyr Cunha Filho, Renisson Neponuceno de Araújo Filho, Ana Luíza Xavier Cunha, Victor Casimiro Piscoya, Guilherme Rocha Moreira, Iloane dos Santos Lima, Ronaldo Dionísio Da Silva, Rejane Magalhães de Mendonça Pimentel, Dayane de Souza Lima, Josué Luiz Marinho, Ricardo Oliveira Silva, Tatijana Stošić, Mílton Marques Fernandes, João Lucas Aires Dias,

Edible Oils Quality and Analysis

Multivariate statistical methods can contribute significantly to classification studies of extra virgin and common olive oil groups. Therefore, nuclear magnetic resonance (NMR) was used to discriminate olive oil samples, multivariate statistical techniques Principal Component Analysis - PCA, Fuzzy Cluster, Silhouette Validation Method to describe and classify. The groups' distinction into organic and common was observed by applying the non-hierarchical Fuzzy grouping with a distinction between the two groups with a 65% confidence interval. The validation was performed by the silhouette index that presented S (i) of 0.73, which showed that the adopted grouping presented adequate strength and distinction criterion. However, PCA only analyzed the behaviors of data from extra virgin olive oil. Thus, the Fuzzy clustering method was the most suitable for classifying extra virgin olive oil.