Graphical analysis of correlations between levels of consumer acceptance, longevity and fragrance silage, through artificial neural networks and analysis of main components

• Autor(a) principal: Matos, Vanessa Gomes
• Data de Publicação: 2020
• Outros Autores: Vianna, Regina Ferreira, Leite, Diego de Jesus
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Research, Society and Development
• Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/3192
• topic: Perfumes; Fragancias; Redes neuronales artificiales; Indicadores de rendimento; PCA.; Perfumes; Fragrances; Artificial neural networks; Performance indicators; PCA.; Perfumes; Fragrâncias; Redes neurais artificiais; Indicadores de desempenho; PCA.

The fragrance and flavoring industry (F&F) generates millions of dollars worldwide and is responsible for the olfactory characteristics of personal hygiene products, perfumes, cosmetics, house holding and any and all products that contain an artificially produced aroma. The research and development of new fragrances are concentrated among the 5 largest fragrance houses in the world and for that reason there is a great concern to maintain the secrecy between the creative processes of these companies. This fierce competition market limits the capacity for innovation in the creation of new products to what is usually successful in the market and the acceptance statistics of those who have been proposed among competitors. These companies are increasingly restricted to innovating within a previously launched universe, producing flankers - versions of products already established in the market. Thus, this article aims to graphically analyze data from a virtual perfumery library, modeled using a multi-layered neural network and resilient backpropagation neural network, validated through principal component analysis. Graphical analysis provides an interpretation of the correlations between levels of consumer acceptance for a perfume and the performance indicators for that fragrance. This study reiterates the existence of correlations between the user's consumption profile and the properties of the fragrances, supporting future studies of exclusive formulation of individually customized compositions for groups or individuals, demonstrating potential use in perfume engineering.