On the structure of classifications recognizable by the pattern recognition algorithm "Core-n"


V.A. Zheligovsky and S. Schwarz


Research Report 97-2
To appear in Computational Seismology
Date: 4 February 1997


The pattern recognition algorithm "Core-n" has been successfully applied in different areas ranging from geophysics to social sciences. We present results on the geometry in the object space of classifications recognizable by the algorithm, including those obtained in computer experiments with low-dimensional problems.

Key phrases

pattern recognition. stability of recognition. Core-n.


The paper is available in the following forms:
1997-2.ps.gz (101kB) or 1997-2.ps (333kB)

To minimize network load, please choose the smaller gzipped .gz form if and only if your browser client supports it.

Sydney Mathematics and Statistics