MULTI-BLOCK DATA ANALYSIS USING COMDIM FOR THE EVALUATION OF FLOURS BY NIR SPECTROSCOPY AND TEXTUROMETER
Flours with different functional properties are being introduced in the human diet. To better know about different some of them, 138 flours samples were submitted to near infrared (NIR) spectroscopy and evaluated in texturometer to check about its hardness, fracturability, adhesiveness, springiness, chewiness, gumminess, cohesiveness and resilience. These measurements were analyzed via Common Dimensions (ComDim) method to verify similarities with properties of the white wheat flour. ComDim is an unsupervised pattern recognition method which is able to extract information from the results that could be presented in informative graphs showing the relationship between the samples through the scores matrix, allowing to check if there is predominance of information in a particular technique (NIR or texturometer in this case) and the contribution of the variables in each technique, which are responsible for the similarities observed in the samples, that can be seemed at the loadings plots. It was possible to infer similarities and differences among the samples studied according to the specific absorption in the NIR region, as well as their texturometer results. The proposed methodology demonstrates that NIR absorption is correlated to hardness and springiness results obtained from texturometer. The use of ComDim to characterize samples demonstrates that properties of hardness and springiness observed in white wheat flour are similar in flours of white corn, oats, rye, white rice and integral rice, golden linseed, banana, sweet potatoes, eggplant, ginger, yacon, açai and amaranth. The proposed methodology was able to discriminate samples that can provide characteristics of hardness and springiness which are similar to the observed for white wheat flour, suggesting that the white wheat flour could be replaced to another with functional properties.