|
|
NEW TASTEL OPERATIONS
|
TASTEL+
|
Before to detail the main innovations developped recently in TASTEL, we can recall the 2003 creation
of a very complete tool: TASTEL+
|
|
TASTEL+ is coming from the gathering of our three previous systems: Tastel, Constel, and Formulatel.
|
|
So, TASTEL+ is designed to cover the three fields of our previous systems:
|
Sensory Analysis,
|
|
Consumer Testing, and
|
|
Formulation optimisations.
|
|
|
|
|
These three operations are working together in order to suggest, if needed, statistical links between
descriptive analysis (Sensory Product Profiles, for example) with preference tests with Preference Mapping
techniques using the Ideal Product modelling technique.
|
|
Or also, the System can suggest ideal formules using existing links between recipes and related
sensory descriptive analyses.
|
|
|
|
The main newest techniques developped in TASTEL are the following ones:
|
|
|
SOUNDS & VIDEOS PROFILES
|
|
It is possible during Sensory Product Profiles or during Consumer Testing to
insert sounds, images, or even videos in the stage of the evaluation.
|
|
This can allow to compare packagings, for instance to test visual impression when walking
along supermarket show-cases, but also to assist the panellist during his evaluation.
|
|
|
|
FLASH PROFILING
|
|
Flash Profiling is a new technique created in ENSIA of MASSY (Essonne/France) by J.M. Sieffermann
to discriminate numberous products.
|
|
This technique has an advantage to rely immediately on the existing sensory knowledge of the experts ;
These ones could be wine-makers, "Noses" in the Parfumery industry or recipe developers.
No sensory training is needed because these experts will define themselves the sensory attributes
which will be the most appropriate to well discriminate the product to assess.
|
|
This technique relies also on a ranking evaluation to avoid product group notation which could be favoured
by using a more classical scoring by intensity marks.
|
|
|
|
Calculations will then give very detailled graphs with a very rich information given by the experts
to discriminate products
More complex calculations are also possible based on STATIS technique (equivalent to Generalized
procustean Analysis) allowing to underline the most common notation characteristics shared by all
the experts in the product scoring ;
|
|
To this end, this technique transforms data by mathematical
techniques as translations, rotations, and similarities.
|
|
Options are suggested in TASTEL to minor the problem of experts using different attribute number
to discriminate products.
|
|
|
|
MAPPING PROFILE
|
|
This very new profile is issue from works coming from the Rennes University, in particular under
the guidance of Pr. Jérôme Pagès.
|
|
Data entry lies to put down products on a grid based on a principle of proximity.
The number of product can be relatively high: 20, 30, up to 50.
|
|
These distances in the plan can be visual, olfactory or gustatory, or general, and these
locations can be justified by open comments entered with a pop-up menu.
|
|
This method is very appropriate in the wine industry where product samples are very numberous, and
where the blending principles are quite very closed of this way of thinking and sorting.
|
|
|
|
Calculations can add the maps of each experts to issue a panel map and a distance matrix between
each product.
|
|
|
|
DTS PROFILE (Time Intensity profile)
|
|
This method
created by Pascal Schlich called Sensation Temporal Dominance (Dominance Temporelle
des sensations) has for aim to study more naturally the modifications of the perceptions.
|
|
The expert can select from 5 to 10 attributes, for example the descriptor which is changing at
this moment in order to score its intensity, and then to select another descriptor in terms of the
change of perception to score this new one, or also to change only the intensity, and so one.
|
|
|
|
This data entry is nearer of the reality than the classical Time intensity technique with
only one attribute, and then, it is possible to issue results based on different parameters:
|
Appearance time (T), which is the time between the beginning of the experiment
and the first selection time of this attribute.
|
|
Scoring (S), which is the intensity mean of this attribute during the experiment, and
|
|
Dominance time (D) which is the time when this attribute was the highest in intensity during
the evaluation.
|
|
|
Principal Component Analyses will then express these results with summary graphs.
|
|
|

Copyright(c) 2005 ABT Informatique. bertrand.thuillier@wanadoo.fr
|