Intensive Course

Intensive Course

Assessment of classification models for medical applications

August 26, 2009 (9:00-13:00)
Lecturer: Vlad Popovici - Swiss Institute of Bioinformatics, Lausanne, CH
Location: Room 40, 28 Baritiu street, Faculty of Automation and Computer Science, Cluj-Napoca

With the advent of new medical technologies, the range of diagnostic and prognostic tools increased considerably. Proteomics, genomics, various imaging modalities are only a few of the approaches generating extremely rich data, on which complex models are built. Before such models are to be used in clinical practice, they need to pass a series of stringent tests and their performance must be correctly assessed.

The purpose of this seminar is to present the most important issues that arise in assessing such models and to provide the statistical tools for answering them. While the main focus will be on the evaluation of classification models in clinical context using the Receiver Operating Characteristics curve, the techniques that will be presented are generally valid and could be applied to any
classification problem.

Biography: Vlad Popovici

Vlad Popovici
Vlad Popovici is researcher with Swiss Institute of Bioinformatics (SIB), working on different theoretical and applied aspects of statistical pattern recognition. He has obtained his Engineering and M.Sc. degrees in computer science from Technical University of Cluj-Napoca, Romania, in 1998
and 1999, respectively, and his Ph.D. from Swiss Federal Institute of Technology (EPFL) in 2004.

Before joining SIB he has worked for Digital Media Institute, Tampere University of Technology as research scientist and for Signal Processing Institute at EPFL as assistant researcher and, later, as postdoctoral researcher. His current research focuses on various issues of machine learning like sparse classifiers, multiple classifier systems, error estimation in small sample size settings and their applications to life sciences.