(Demande de profil Ing 351nieur d'Etude ACS 18.10.10 -2)

Vdeclercq

Informations

• Publié par: Kivim
• Nombre de lectures: 24
• Langue: English

Extrait

REFERENCES

Corps

: Ingénieur d’Etude

Bap n°

: A

N° de l'emploi-type

:

n°A2A21

Intitulé de l'emploi-type

: Ingénieur en traitement de données biologiques

Centre de recherches

: INRA Clermont Theix 63122 St Genès Champanelle

Intitulé de l'Unité

: Unité de Recherche sur les Herbivores - Equipe Adaptation et

comportements sociaux

Personne(s) à contacter

: Isabelle Veissier

Téléphone :

Adresse électronique :

__________________________________________________________________________

_

Définition de l’emploi :

L'ingénieur en traitement de données biologiques organise la

collecte et réalise la gestion et le traitement de données issues de la recherche en sciences

du vivant

Environnement et contexte de travail

(lieu d'exercice, conditions de travail particulières,

contraintes) :

L’ingénieur participera au projet Welfur qui vise à développer des outils d’appréciation du

bien-êrtre des animaux à fourrure (vison, renards). Ce projet est porté par l’Efba (European

fur breeders association)

.

Activités :

Welfare is a multidimensional concept, embracing physical and mental health. Welfare

Quality® proposes 12 criteria that must be fulfilled In order to respect animal welfare:

Absence of hunger, Absence of thirst, Comfort around resting, Thermal comfort, Ease of

movements, Absence of injuries, Absence of diseases, Absence of pain due to management

procedures; Expression of social behaviour, Expression of other behaviours, Good human

animal relationship, Positive emotional state.

These criteria can be grouped into 4 principles:

Good feeding, Good housing, Good health, Appropriate behaviour. The assessment of

welfare requires numerous measures to be performed on farms. Welfare Quality

®

designed

a formal model where data on animals or their environment are transformed into value

scores that express compliance with the 12 criteria then the 4 principles. Finally farms are

sorted in four categories: excellent welfare, enhanced welfare, basic welfare and not

classified. This evaluation model is tuned according to experts from animal and social

sciences, and stakeholders (Figure 1).

Figure 1: Welfare Quality

®

hierarchical model of evaluation. On Step 1, animal scientists who

developed welfare measures were consulted. On Step 2, Welfare Quality® consulted animal

(for their knowledge on animals) and social scientists (for their knowledge on societal

expectations). On Step 3, these people and stakeholders likely to use the evaluation system

were consulted. Experts were asked to give their opinion on datasets and the model is

designed so as to best match these opinions.

30-50

measures

12

criteria

Overall assessment

4

principes

Step 1

Step 2

Step 3

30-50

measures

12

criteria

Overall assessment

4

principes

30-50

measures

12

criteria

Overall assessment

4

principes

Step 1

Step 2

Step 3

On Step 4, a software can be developed to ease the calculation of scores. The software is

internet based. It includes a database of data collected on farms, a module to calculate

scores and the scores are in turn stored in the database. A template is designed for users to

load their data. Access rights are given so that users have access to their own data.

Statistics are produced (e.g. distribution of scores across al farms monitored) and available

to all public, with no reference to individual information. A module allows simulating

improvements on farm and seeing their impact on the scores. The application developed for

species covered by Welfare Quality

®

can be visited at http://www1.clermont.inra.fr/wq/.

Work proposed in Welfur

We propose to concentrate on Step 1 of the construction of the evaluation models for minks

and foxes (two species). The following activities will be carried out to complete Step 1

1- To liaise with experts developing the measures to check welfare criteria in order to

decide what measure should be used in scoring and to specify – when necessary –

the way the measures are taken and the type of data they produce.

The two experts

leading foxes and minks are respectively Steen Henrik Moller (Aarhus University-DK)

and Jaakko Mononen (Kuopio University-FI). Close links with them will be

established.

2- To propose ways to integrate the information at criterion level. At that stage we

define the type of formula that will be used to combine the measures (eg decision

trees, weighted sums, comparison to thresholds, spline functions, Choquet integrals).

The type of formula is chosen according to the number of measures to be used in a

given criterion and to their nature (ordinal, cardinal, qualitative; taken at animal or

group level etc.). at that stage only the type of formula is decided.

3- To prepare the consultation of experts. Virtual datasets are prepared for experts to

react on them. They are constructed in a way to assure that enough information will

be obtained to be able to configure properly the formula defined previously. The

datasets are presented in excel file where macros that will help experts to express

step by step their opinion on each set of data (eg alternatives including % moderately

injured and % severely injured animals are shown, the experts first rank them then

the alternatives are sorted according to their rank and finally experts attribute

scores). A list of experts will be provided by EFBA.

4- To run the consultation with experts. The excel tables may be sent by e-mail. We feel

it is better if we can have a face-to-face meeting with all experts involved in a given

species. Experts can fill in the tables on their own then discuss together and

exchange opinion and finally refine (or not) their answers.

5- To process the answers of experts. This includes calculation of means, estimation of

weights for weighted sums and spline finctions.

6- To present the results in a concise and practical way. For each criterion, the

measures used will be clearly set, and the formula will be explicit so that a third

person can calculate welfare scores.

Step 2: Step 2 uses a specific operator, the Choquet integral, which allows attributing

intrinsic weights to criteria and to limit compensation between them. In Welfare Quality

®

Step

2 of the evaluation model differs very little across species. That is to say, the importance to

be attributed to various criteria and their possible compensation (or not) was similar between

species. We thus propose to work on the data collected in Welfare Quality® from the

consultation run in Step 2 and to extract average values for the various weights to be used in

the construction of principles.

L

’ingénieur recruté développera ces étapes sous la supervision d’I.Veissier.

Compétences :

Connaissances généralistes sur l’animal

Rédaction

Anglais courant

Goût pour les mathématiques, la formalisation des conn

ai

ssances

Diplôme réglementaire minimum exigé :

Ecole d’ingénieur agronome ou agricole ou équivalent (Bac +5)

Formation recommandée :

Ecole d’agronomie, master 2

Capacités personnelles :

Intérêt pour le bien-être animal

Esprit logique et synthétique

Sens de l'organisation

Capacité à interagir avec des spécialistes du domaine

Contrat à Durée Déterminée

d’une durée de 9,5 mois

Date de prise de fonction :

18/10/10

Salaire

: 1917,00 €

Brut