Discriminação de erva-mate para chimarrão quanto à origem geográfica e presença de açúcar utilizando FTIR e quimiometria
Ilex paraguariensis or yerba mate is a commercially important plant in the food sector, especially chimarrão, a typical drink from South Brazilian states – Paraná (PR), Santa Catarina (SC) and Rio Grande do Sul (RS) – the largest national producers. Like wines and other products, yerba mate changes...
| Autor principal: | Lima, Patricia Casarin de |
|---|---|
| Formato: | Trabalho de Conclusão de Curso (Graduação) |
| Idioma: | Português |
| Publicado em: |
Universidade Tecnológica Federal do Paraná
2020
|
| Assuntos: | |
| Acesso em linha: |
http://repositorio.utfpr.edu.br/jspui/handle/1/6753 |
| Tags: |
Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!
|
| Resumo: |
Ilex paraguariensis or yerba mate is a commercially important plant in the food sector, especially chimarrão, a typical drink from South Brazilian states – Paraná (PR), Santa Catarina (SC) and Rio Grande do Sul (RS) – the largest national producers. Like wines and other products, yerba mate changes chemically and sensorially according to the region of cultivation, which encourages the practice of adulteration, such as adding sugar to mask unwanted sensory aspects or misleading provenance to add false value. The objective of this work was to use FTIR spectroscopy, alternatively to standard methods, combined with chemometrics in the classification of yerba mate as origin and presence of sugar, for possible application in the control of adulteration. The KBr pellet technique was adopted to support the samples. Duplication assays were performed on the Shimadzu IRAffinity-1 spectrometer for the 69 samples from the three southern states. In the Matlab R2008b, the spectra were pretreated, including variable selection, working with the information of real importance. From the duplication, the average spectra were extracted, for which the PLS-DA classification method was used. A good discriminating model was obtained about the origin of the yerba mate using 15 latent variables. The sensitivity was higher than 62% and selectivity greater than 89% for the three classes. After cross-validation, the AUC was 0.9133. For the presence of sugar, the model with 10 latent variables was chosen, whose accuracy was 65%, sensitivity 62% and 75% for the classes without sugar and with sugar, respectively, and AUC 0.6923. Therefore, even considering the high amount of latent variables required, the use of PLS-DA in FTIR spectra was satisfactory, considering the complexity of the studied matrix and the limitations of a linear method. |
|---|