|
IDeAL Research Group |
|
|
Our research at the Intelligent
Decision and Data Analysis Lab (IDeAL) focuses on the development of new methodologies for the data mining of large-scale object-relational databases. We apply the end results
of our research within the Anthropometry, BioInformatics
and Health Care domains.
Contact:
Please contact us at hlviktor{at}site.uottawa.ca if you are
interested in our research.
Current projects
include the following:
-
Mining large-scale relational databases. Relational data
mining refers to the problem setting where data resides in multiple tables
(or relations) as contained in a
relational database. Researchers and practitioners agree that this field is of
strategic importance due to the vast amounts of real world data that is currently
stored in this format. Consider a database containing Terabytes or Petabytes of data. In this case, the evaluation of a
hypothesis may involve hundreds of thousands of tuples
spread over multiple tables, leading to computationally expensive multiple joins,
which cannot assume the use of main memory. Furthermore, the current
state-of-the art, involve object-relational databases which contain also multimedia content such as 2D images or
3D objects. We have developed the so-called IDeAL2 utility-based environment to directly mine data as contained in medium-sized
object-relational databases, focusing on techniques for classification and clustering. Currently, we are extending this work to address very-large scale
databases.
-
Finding clothes that fit.
In the apparel
industry, an important challenge is to produce garments that fit various
populations well. However, repeated studies of customers’ levels of
satisfaction indicate that this is often not the case. The following questions
come to mind. What, then, are the typical
body profiles of a population? Are there significant differences between
populations, and if so, which body measurements need special care when
e.g. designing garments for Italian females? Within a population, would it be
possible to identify the measurements that are of importance for different
sizes and genders? Furthermore, assume that we have access to an accurate anthropometric database. Would there, then, be a
way to guide the data mining process
to discover only those body
measurements that are of the most
interest for apparel designers? To this end, we are investigating new
approaches to explore a database, containing anthropometric measurements and 3-D body scans, of samples of the
North American, Italian and Dutch populations.
-
Drug design, protein interaction and the docking problem. We have developed the Capri/MR system which makes it possible to
retrieve proteins of similar three-dimensional shape, as contained in very
large protein structure databases such
as the Protein Data Bank (PDB), which contains around 55,000 different protein
structures. The main applications of our system are in structural proteomics,
protein evolution and mutation and drug design, in particular for the computer
aided design of non-toxic drugs. Currently, we are studying the use of data mining and computational
intelligence techniques for protein
family prediction and aim to address the docking problem.
The following past projects has been completed successfully.
-
Preserving software dependent data
over a very long time. The rapid changes in technology in general, and in
Internet-related technologies in particular, make the long-term preservation of
e-data an important challenge. Our
objective was to better understand the intrinsic subtleties when preserving
e-data over 50 years or more. To
this end, our research aimed to creating an environment to study the long-term
preservation of e-data. We focused our attention on preserving multimedia and
relational data, which were dependent on software components, for future use.
The end result of this research resulting in the IDeaL
long-term experimental environment, containing a persistent data webhouse, together with archiving and indexing, retrieval
and trend analysis modules for handling the evolving e-data.
-
Managing and exploring Cultural
Heritage repositories. We studied
the efficient management and exploration of very large repositories of 2D
images and 3D objects for the modelling and
reconstitution of complex heritage sites, and applied our
methodology to a variety of real cases.
Collaborators and Sponsors:
-
National Research Council of
-
-
IBM Canada
-
-
-
-
National Science and Engineering
Research Council (NSERC) of