Redes neurais artificiais: nova abordagem para identificação da qualidade da bebida do café arábica utilizando-se resultados de análises químicas

• Autor(a) principal: Messias, José Américo Trivellato
• Data de Publicação: 2009
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: LOCUS Repositório Institucional da UFV
• Texto Completo: http://locus.ufv.br/handle/123456789/672
• topic: Rede neural artificial; Qualidade do café; Classificação; Artificial neural network; Quality of coffee; Classification; CNPQ::CIENCIAS AGRARIAS::ENGENHARIA AGRICOLA::ENERGIZACAO RURAL

The present study was performed with the objective of evaluating the utilization of an artificial neural network to correlate results from sensorial analysis with the chemical analysis of coffee samples. The coffee samples utilized belonged to the coffee plant Coffea arabica L., cultivars Acaiá do Cerrado, Topázio, Catuaí vermelho 99, Acaiá 474-19 and Bourbon, collected in the southern region of Minas Gerais State, Brazil, with controlled post-harvest processing. Performed chemical analyses were those of non-reducing sugars, reducing sugars, total titratable acidity, electrical conductivity, leaching of potassium and polyphenols. Coffee quality was evaluated by the sensorial analysis. The multiple linear regression fits of the experimental data from sensorial analysis as a function of the data from the chemical analysis presented a determination coefficient of 0.6681, while the employed artificial neural network reached a classification success rate of, approximately, 85% for the values of sensorial analysis. The artificial neural network obtained was tested with coffee samples collected from properties in the same region, with uncontrolled postharvest processing, reaching a 75% success rate for drink classification. The obtained result is an indicator that the artificial neural network is a suitable tool to correlate results from sensorial analysis with the chemical analysis.