Identification of Key Textural Attributes of Fluid and Semi-Solid Foods Using Regression Analysis

Abstract

Regression analysis was used to identify key attributes from 15 textural terms generated by a panel for 27 fluid and semi-solid commercial foods. A search using single independent variables showed that “thick” gave the best average R2 with a value of 0.548; a search with two independent variables showed that “thick” and “soft” gave the best average R2 of 0.748; a search with three attributes showed that “slippery”, “thick,” and “soft” gave the highest R2 values of 0.803. The final equations provided a set of regression parameters which can be used to predict twelve textural attributes from scores obtained for “thick”, “soft”, and “slippery”.