Applications of Softcomputing Techniques in Biomedical Domains

Project Members:

• Prof. Dr. Barbara Hammer
• Dipl.-Inf. Alexander Hasenfuss

Partners:

• Dr. Michael Biehl, University of Groningen, The Netherlans
• Dr. Thomas Villmann, University of Leipzig, Germany
• Dr. Frank-Michael Schleif, University of Leipzig, Germany
• Dr. Marc Strickert , IPK Gatersleben, Germany
• Dr. Peter Tino, University of Birmingham, UK

Duration: since 01/2005

Project Description:

The amount of electronic data in biomedical domains such as medical images, mass spectrometric profiles, or SNP data are dramatically increasing so that automated data mining tools constitute an indispensible technique in these domains. The application domain puts particular challenges on machine learning methods such as extremely high dimensionality and only sparse data, imbalanced class distributions, etc. We developed a variety of highly flexible classification and data inspection tools which have been applied to diverse areas including proteomic profiling and microarray analysis. In this context, two (partially funded) seminars (a Dagstuhl seminar and a MPIPKS seminar) gathering together experts in related fields were co-organized by Prof. Hammer.