SVM-Theory - Support vector machines and kernel methods : theory, algorithms and applications

From 01-01-2002 to 31-12-2005

Description

SVM and kernel based methods are promising techniques for the future. This follows from our own research experience and recent attention and impact of these methods in several areas such as neural networks, datamining machine learning, signal processing, statistics, circuits and systems theory, system identification, control theory and mathematics. The present project aims at further investigating these methods concerning fundamental theoretical aspects, design of efficient and reliable algorithms and applications ro real-life problems. We plan to study the following themes : (1) LS-SVM extensions, fundamental theory, links with several methods in different areas (2) statistical analysis (3) large scale datamining algorithms and optimization (4) systems and control theory (5) bio-informatics (6) biomedical application (7) adaptive signal processing. The SVM research expertise has been built up at ESAT/SISTA under the supervision of J. Suykens, B. De Moor and J. Vandewalle. The resesearch expertise of S. Van Huffel and M. Moonen on biomedical applications and adaptive signal processing, respectively, is complementary in this study. As objective we pose that thise interdisciplinary approach will create a strong interaction between the theoretical analysis and its application in the above mentioned areas.

Team

• Johan Suykens, Coordinator
• Marc Moonen, Promoter
• Bart De Moor, Promoter
• Sabine Van Huffel, Promoter
• Joos Vandewalle, Promoter
• Kristiaan Pelckmans, Team member
• Marcelo Rodrigo Espinoza Tapia, Team member
• Samuel Xavier de Souza, Team member
• Carlos Andres Alzate Perez, Team member

Financing

Funding: FWO - Research Foundation - Flanders

Program/Grant Type: FWO Research Grant - FWO Research Grant