Análise sequencial no ajuste de mapas de preferência

• Ano de defesa: 2018
• Autor(a) principal: Fagundes, Cássia De Souza Santos
• Orientador(a): Ferreira, Eric Batista
• Banca de defesa: Oliveira, Marcelo Silva De, Gomes, Davi Butturi
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Alfenas
• Programa de Pós-Graduação: Programa de Pós-Graduação em Estatística Aplicada e Biometria
• Departamento: Instituto de Ciências Exatas
• País: Brasil
• Palavras-chave em Português: Análise de componentes principais; Estatística
• Área do conhecimento CNPq: PROBABILIDADE E ESTATISTICA::PROBABILIDADE E ESTATISTICA APLICADAS
• Link de acesso: https://repositorio.unifal-mg.edu.br/handle/123456789/1312

The External Preference Map (EPM) is one of the statistical tools applied to the sensory analysis of food and beverage, being widely used to identify the acceptance of consumers in relation to the products evaluated. In summary, it can be understood as the overlap of response surfaces adjusted to the consumer acceptance data according to the sensory attributes of a product. In the literature that bases the EPM, a sample size of at least 100 consumers is recommended to evaluate the acceptance of the products. One way of not prefixing sample size is to use sequential tests that also control the error rates type I and type II. One manner to use these EPM tests is to infer the quality of their adjustment. Thus, the objective of this work is to propose a sequential approach in the construction of EPM, infer in its quality of adjustment by means of the coefficient of determination (R2) and propose a way to select models sequentially. The analyzed data were obtained through a sensory analysis of the guarana flavor soft drink from two brands in two versions (traditional and zero sugar). The adjusted sequential preference map obtained a minimum adjustment of 70% by the sequential T test and was built with 40 consumers. Thus, it was possible to observe that the traditional version of both brands was better accepted, but the market leading brand stands out in the average acceptance. In the comparison of models, among the models existing in the literature, the vector was the one that presented the best fit. Finally, it is concluded that it is possible to decide sequentially about the quality of preference maps adjustment, so that it can be constructed without a prior determination of the number of consumers to be interviewed.