2.3.2. Comparison of the IdMap and the PrefMapD
Although both methodologies (IdMap and the PrefMapD) are
similar, the nature of the data involved differs. And it is especially
according to this difference that both techniques are compared.
2.3.2.1. Link between the maps (a priori). According to Worch et al.
(2012), a consumer is consistent if the ideal product provided
shares similar characteristics with the preferred product.
Thus, in the sensory space, the projection as illustrative of the
ideal profiles provided by a consumer must be close to the preferred product. To check that, the corrected ideal profiles of the
consumers (Table 4b) on one hand and the hedonic scores on the
other hand (Table 4c) are, respectively, projected as supplementary
entities and supplementary variables on the sensory space (Table
4a). For a panel composed of consistent consumers, the distribution of the consumers within the correlation circle (hedonic) is
coherent (i.e. going in the same direction) with the distribution
of the projections of their ideal profiles within the sensory space.
This double projection provides a first visual impression about
the relationship between IdMap and PrefMapD. However, it is only
approximate as the one to one relationship is not further studied
here.
2.3.2.2. Liking threshold and definition of the PrefMapD and IdMap
maps. In this study a product is defined as liked by a consumer if
the consumer considered would like it more than a certain threshold value. Similarly, each consumer is associated to an area of liking
corresponding to the area of the sensory space which contains
products he/she would like.
The construction of the PrefMapD depends on the threshold value (also called liking threshold), as it determines whether a product
belonging to the sensory space would be liked or not by each
consumer.
In practice, for a given consumer, each point of the space is
associated with a product whose liking score is estimated based
on a model defined for that consumer. This estimation of the liking
score is compared to the liking threshold: if the estimation is larger
(resp. smaller) than the threshold, the product corresponding to
this point of the space is liked (resp. rejected) by the consumer.