Development of a Model for Prediction of Consumer Liking from Visual Attributes of New and Established Apple Cultivars

Abstract

Consumer liking of the appearance of new and established apple cultivars was determined on an hedonic scale. Descriptive analysis, with a panel of 18 judges, was used to characterize the appearance of 10 red apple cultivars which varied in overall size, taper, ribbing, symmetry, uniformity-of-shape, length/width ratio, degree of striping, percent overcolor, brightness of red over-color, groundcolor and uniformity-of-color. Stepwise multiple regression was used to develop models to predict liking from the visual attributes. A one-variable model with extent of ribbing predicted liking, for an experienced panel while a one-variable model with degree of apple taper predicted consumer liking. The effectiveness of the consumer model was validated with a second consumer evaluation, on collections of red or yellow apples with varying taper values. While the developed model was not able to predict the actual liking score, it was able to approximate the rank liking of the yellow apples.