Page 363 - Special Topic Session (STS) - Volume 4
P. 363
STS1080 Karina G. et al.
in his menu, even in he dislikes, when there is no other alternative. This will be
internally managed by introducing penalization or bonification of associated
dishes.
2.5. Cultural Eating Styles
There is a cultural factor in the nutrition habits of a person according to
where the person lives. For instance, in a Mediterranean country, the breakfast
often is more caloric and less proteic, the lunch concentrates more proteins,
and dinner can concentrate more vegetables or fruits. In the Anglo-Saxon-
style breakfast tends to be more proteic, lunch is light with vegetables and not
much calories, and dinner more proteic. Diet4you can manage this contextual
knowledge in the composition of menus by getting Diet-Styles in form of
tables with probability distribution of food or nutrients families conditioned
to a variable number of meals per day. From this information the general
nutritional plan is divided into sub-plans for each specified meal (breakfast,
lunch and dinner for example). The personal menu planer is building local
menus for each meal and guarantees that the whole resulting menu fits the
global nutritional plan originally prescribed.
2.6. Personal Menu Planner
The personal menu planner (PMP) is mainly implemented following the
cycle of Case-Based Reasoning. Given a nutritional plan for a certain individual
i: νi = <Fi, Ti, Qi>, and considering that the Fi vector contains the N families of
food resulting from a certain level of granularity determined in the reference
food ontology:
1. Pre-processing step. Pre-process the DB in order that all food families
have equivalent units in Kcal and build the transformed data base Food
Proportions Data Base (FPDB). The FPBD contains either prepared
dishes or simple foods d with
• = ( , … , ), being the proportion of food family
1
contained in one standard portion of dish ( = 1: ).
• is the quantity associated to one standard portion of dish d, in
grams or cups or the corresponding measurement unit.
2. Retrieval step. pd is a vector of proportions and thus, it is directly
comparable with Fi. .The Euclidean distance is suitable to compare
composition of two dishes through their pd . FPDB is used as the case
base to identify candidate dishes with pd close to Fi prescribed in the
∗
targeted nutritional plan. Sort the elements in FPBD into =
{ () | () ( () , ) ≤ ( (+1) , ). Candidates will be
in the first positions of and recommended menu is composed
∗
by using iCG strategy (iterative Candidates Generation) validated in
previous works [CCIA 2017, CCIA 2018]. At each iteration, the candidate
352 | I S I W S C 2 0 1 9