RESUMO
BACKGROUND: Brazil has consolidated a relevant position in the world market, being the largest exporter and second producer of beef. Genetics, feeding system, geographic origin and climate influence the multielement profile of beef. The feasibility of combining classification algorithms with major and trace elements was evaluated as a tool for authentication of beef cuts. METHODS: Animals of Angus, Nelore and Wagyu crossbreeds, raised in a vertically integrated system, were sampled at the slaughterhouse for chuck steak, rump cap and sirloin steak. Supervised learning algorithms i.e. Classification and Regression Tree (CART), Multilayer Perceptron (MLP), Naïve Bayes (NB), Random Forest (RF) and Sequential Minimal Optimization (SMO) were used to build classification models based on the multielement profile of beef determined by neutron activation analysis. RESULTS: Br, Co, Cs, Fe, K, Na, Rb, Se and Zn were determined in the beef samples. The classification accuracy values obtained for the beef cuts were 96% (MLP), 95% (SMO), 91% (RF), 86% (NB) and 70% (CART). CONCLUSION: The Multilayer Perceptron algorithm provided the best classification performance towards authentication of beef cuts on basis of major and trace element mass fractions.
Assuntos
Algoritmos , Aprendizado de Máquina , Animais , Bovinos , Teorema de Bayes , Algoritmo Florestas Aleatórias , BrasilRESUMO
Brazilian livestock with a herd of more than 215 million animals is distributed over a vast area of 160 million hectares, leading the country to the first position in the world beef exports and second in beef production and consumption. Animals risen in the biomes Amazônia, Caatinga, Cerrado, Pampa and Pantanal were selected for this study. Beef samples were analyzed for their elemental content by neutron activation analysis and classified according to their origin by three machine learning algorithms (Multilayer Perceptron, Random Forest and Classification and Regression Tree). Significant differences (p < 0.0001) were observed between the beef elemental content from the different biomes for all multivariate contrasts using NPMANOVA. The highest classification performance was obtained for the biomes Amazônia and Caatinga using Multilayer Perceptron. Results showed the feasibility of combining trace element content and machine learning approaches for the Brazilian beef traceability.
Assuntos
Aprendizado de Máquina , Redes Neurais de Computação , Carne Vermelha/análise , Oligoelementos/análise , Animais , Brasil , Bovinos , Ecossistema , Carne Vermelha/classificaçãoRESUMO
Epífitas são eficientes indicadores de poluição atmosférica devido à absorção de elementos químicos diretamente da atmosfera. Folhas de onze espécies de bromélias e uma espécie de orquídea foram coletadas no Parque Estadual Carlos Botelho, SP, para a determinação de dezesseis elementos químicos. A seleção foi baseada na diversidade e acumulação de elementos químicos encontrados nas folhas. Com um índice de acumulação de 0,7, a espécie Canistropsis billbergioides tem potencial para ser empregada como biomonitora de elementos químicos na Mata Atlântica.
Epiphytes are efficient indicators of atmospheric pollution because of their direct uptake of chemical elements from atmosphere. Leaves from eleven species of bromeliads and one species of orchid were collected in the Parque Estadual Carlos Botelho, SP, for determination of sixteen chemical elements. The selection of the best biomonitor species was based on the diversity and accumulation of chemical elements in the leaves. With an average accumulation index of 0.7, Canistropsis billbergioides has potential to be used as a biomonitor of chemical elements in the Atlantic Forest.