<p>The application of AI systems to address the complexity and limitations of sensory evaluation is gaining interest in food science research. The aim of this study was to analyse the sensory evaluation data of gluten-free cakes containing different proportions of legumes and pseudocereals using neutrosophic similarity measurement and to determine the effects of legume and pseudocereal flours on cake quality parameters. In this context, the sensory analysis results of cakes prepared by adding chickpea, buckwheat and quinoa flours to rice flour in three different proportions (10, 20, 30%) were evaluated using a fuzzy logic application based on neutrosophic similarity measures (Jaccard and Dice). Both models identified the most and least preferred samples by the panelists and the samples suggested by the models and the samples closest and farthest from the model were compatible with each other. The cakes with the highest neutrosophic similarity score were those containing 30% quinoa flour, 10% mixed flour and 20% mixed flour, while the cakes with the lowest neutrosophic similarity score were those containing 20% and 30% buckwheat flour and 30% chickpea flour. The physical, chemical, textural, functional and nutritional properties of the cakes and their pGI values were also determined. The addition of these flours increased SDS content and antioxidant capacity, while decreasing RDS content, SHI and pGI values. This study shows that neutrosophic sets, a special case of fuzzy logic, can be used in the food industry and proposes an alternative method using machine learning instead of complex sensory panels involving human factors.</p>