Workpackages short description
Workpackage WP1 is dedicated to the management of the project.
WP2 User Requirements and Evaluation
The aim of workpackage WP2 is
- to define user requirements,
- to suggest treatment and methodology to solve user problems,
- to evaluate continuously the project results,
- to prepare pilot application and evaluate the software
with the benchmark.
External evaluation will be performed by the way of Peer
Workpackage WP3 addresses the problem of constructing the
final software kernel from all the prototype components of
the project. The components will be integrated in the kernel
software: the Workbench. In parallel, software components
will also be available for developers in a library of modules.
In order to answer to user requirements on treatment guidance,
automatic treatments will be chained. A user manual, supervised
by user leader, will be available on a user-friendly display.
WP4 Data Management
Workpackage WP4 is concerned with access to the data by all
the methods. This frame was designed in the SODAS project
but main improvements are needed and are asked by the users.
Effort must be given to:
- Data Format SOM library that interfaces different algorithms
with the data, which must take into account new data format
- Creation of SO from the database,
- SO editing based on the small prototype module developed
in SODAS following the request of users,
- Propagation on the database meaning the retrieve of individuals
corresponding to SO's.
Moreover a metamodel will be designed that will hold information
for the assertions that construct the SO. The data format
(SOM) will have to take the metamodel into account.
WP5 Symbolic Statistics, Dissimilarities
Workpackage WP5 as workpackages WP6 and WP7 is dedicated
to the development of new methods. It concentrates on primary
operations on data. It extends standard statistics to symbolic
data and computes new dissimilarities between SO, with a support
to users in selecting dissimilarity measure for their data
This workpackage provides also a way to select variables
and simplify a set of SO.
WP6 Unsupervised Classification, Validation
and Cluster Representation
WorkpackageWP6 develops new methods for unsupervised classification
like pyramid clustering, partitional clustering and factorial
methods of SOs. Kohonen maps are developed for visualising
SOs. For all clustering methods, a validation procedure is
used to study the robustness of the results.
For more information see http://www-rocq.inria.fr/sodas/WP6/
WP7 Supervised Classification and Validation
Workpackage WP7 develops methods for supervised classification
All visualisation aids have been gathered in workpackage
WP8. Improvement will be added to the actual SODAS representation
of a SO, following users requirements. In particular, it will
be possible to import figures inside other software, with
The outputs of the methods will be visualised in an interactive
way by the means of trees, bi-plot, pyramid, Kohonen graphic
and correlation circle.
WP9 Exploitation and Dissemination
The missions of workpackage WP9 are to prepare the dissemination
plans, to prepare an detailed exploitation plan on the future
exploitation, to establish a network of statisticians, to
establish a user group, to enhance visibility of the project
and create awareness within the target user community (especially
Statisticians from Member States and Central European Countries)
and to prepare schools.