INB-Lunch-Seminar

Advanced methods for prototype-based classification 

Petra Schneider

Classification based on prototypes by means of Learning Vector
Quantization (LVQ) is a particularly intuitive and flexible tool which
has been applied in a variety of areas like biology and medicine. The
approach is especially attractive, since the classification model allows
for an immediate interpretation and provides insights into the nature of
the data and the classification problem. Several variants of LVQ have
been developed recently, of which Robust Soft LVQ is a promising one.
An important ingredient of LVQ systems is the employed distance measure.
The talk presents a new technique for metric learning in LVQ. The method
is illustrated by an application from the medical domain. Furthermore,
modifications of Robust Soft LVQ are shortly presented.