Skip to main content
ARS Home » Research » Publications at this Location » Publication #137850
Title: MODELING CONSUMER RESPONSE TO FRESH TOMATO (LYCOPERSICON ESCULENTUM, MILL.) FLAVOR BY PARTIAL LEAST SQUARES REGRESSION

Author
item Margaria, Carlos
item WEST, JENNIFER - UNIVERISTY OF GEORGIA
item ABEGAZ, EYASSU - UNIVERISTY OF GEORGIA
item TANDON, KAWALJIT - USDA, CITRUS & SUBTROP.
item Baldwin, Elizabeth - Liz
item SCOTT, JAY - UNIV OF FL, GCREC
item SHEWFELT, ROBERT - UNIVERISTY OF GEORGIA

Submitted to: National Meeting of Institute of Food Technologists/Food Expo
Publication Type: Abstract Only
Publication Acceptance Date: 8/1/2002
Publication Date: 9/2/2002
Citation: Margaria, C., West, J., Baegaz, E.G., Baldwin, E.A., Tandon, K., Scott, J.W., Shewfelt, R.L. Modeling consumer response to fresh tomato (Lycopersicon esculentum, Mill.) Flavor by partial least squares regression. Annual Meeting of Institute of Food Technologists. 2002. Abstract p. 178.

Interpretive Summary:

Technical Abstract: Consumer acceptability of tomatoes harvested at maturity stages "red" and "breaker" was modeled as a function of sensory descriptors. Those were then modeled as a function of volatile and non-volatile compounds. The resulting models significantly reduced the number of variables needed to explain the variability, and explained a high percentage of the variability due both to the model effects and the dependent variables. The variable sets selected for each model were combinations of taste and aroma except for one that combined taste and a chemosensory sensation with no aroma component. The sensory descriptors present most frequently in the models were "salty", "green", "over-ripe" and "earthy". Flavor components that were present most frequently in models of descriptors were ethanol, 3-methylbutanol and 6-methyl-5-hepten-2-one. Partial least squares regression provided a useful technique to select variables to build a model for consumer acceptability as a function of a laboratory data.