Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network

  Advanced Search  

PREPRINT. Proc. of the ECML'98 Workshop on UpgradingLearning3 the MetaLevel: Model Selection and

Summary: PREPRINT. Proc. of the ECML'98 Workshop on
UpgradingLearning3 the Meta­Level: Model Selection and
DataTransformation, pages 54­65, 1998.
Model selection using measure functions
Arne Andersson 1 , Paul Davidsson 2 , and Johan Lind’ en 1
1 Dept. of Computer Science, Lund University, Box 118, S--221 00 Lund, Sweden.
2 Dept. of Computer Science, University of Karlskrona/Ronneby, S--372 25 Ronneby, Sweden.
arne@dna.lth.se, pdv@ide.hk­r.se, johan@dna.lth.se
The concept of measure functions for generalization performance is suggested. This
concept provides an alternative way of selecting and evaluating learned models (classi­
fiers). In addition, it makes it possible to state a learning problem as a computational
problem. The the known prior (meta­)knowledge about the problem domain is captured
in a measure function that, to each possible combination of a training set and a classifier,
assigns a value describing how good the classifier is. The computational problem is then
to find a classifier maximizing the measure function. We argue that measure functions are
of great value for practical applications. Besides of being a tool for model selection, they:
(i) force us to make explicit the relevant prior knowledge about the learning problem at
hand, (ii) provide a deeper understanding of existing algorithms, and (iii) help us in the
construction of problem­specific algorithms. We illustrate the last point by suggesting a


Source: Andersson, Arne - Department of Information Technology, Uppsala Universitet


Collections: Computer Technologies and Information Sciences