Application of chemometric tools in the development of food bars based on cocoa shell, soy flour and green banana flour

Abstract

In this study, exploratory methods and the simplex-centroid design were applied in the development of food bars with cocoa shell, soy and green banana flours that were evaluated sensorially. The results of the chemical composition and texture were submitted to the Scott-Knott test and the Principal Component Analysis, and was observed a significant difference between the treatments (P < 0.05), as well as a tendency to form five groups, according to its nutritional characteristics. The design used to optimise the sensorial acceptance of the food bars was effective, whose contour surface indicated an optimal region with greater acceptance, with all formulations having good acceptance (scores above 5.52). Check-All-That-Apply, Contingency Analysis and Hierarchical Cluster Analysis allowed describing the products, demonstrating the sensory differences between treatments with the presence/absence of cocoa shell. It is concluded that the studied ingredients can be used in food bars, providing nutritional and sensory characteristics acceptable to the consumer.

Artificial Neural Networks for the Evaluation of Physicochemical Properties of Carrots (Daucus carota L.) Subjected to Different Cooking Conditions as an Alternative to Traditional Statistical Methods

“…The implementation, training, and data clustering were performed on a computer equipped with Intel Core i5-7200U CPU@2.50 GHz 2.70 GHz based on a ×64 architecture with 8 GB RAM and Windows 10 operating system with MATLAB (version R2019b) software using the SOM Toolbox for MATLAB (Vesanto et al, 1999) . The principle of the technique of SOMs can be found in other studies in the literature. ,,, …”

“…22 The principle of the technique of SOMs can be found in other studies in the literature. 5,6,23,24 The data obtained from the analyses performed were organized in a matrix with a dimension of 18 samples by 12 variables. The size of the network, or the number of neurons, was determined by applying the heuristic formula 5√n (n, is the number of samples).…”

“…Among unsupervised multivariate statistical methods, principal component analysis (PCA) and hierarchical cluster analysis (HCA) have been widely used for quality assessment and fruit and vegetable processing, − so the principle involves classifying and grouping data in order to look for similarities …”