Abstract. The analysis of sequences is one of the major research areas
of bio-informatics. Inspired by this research, we investigate the discov-
ery of sequential patterns for use in classification. We will define varia-
tions of a fit function that enables us to tell if one pattern is "better"
than another. Furthermore we will show how domain knowledge can be
used for faster discovery of better sequential patterns in specific types of
databases, in our case a receptor database.
All algorithms on the website of Edgar de Graaf (http://www.liacs.nl/~edegraaf/) are provide under the GNU GENERAL PUBLIC LICENSE. Please read this license before using the software (gpl.txt).
- Algorithm without Probable Time Window (PTW) and the Small Pattern principle (SP).
- Algorithm with PTW only.
- Algorithm with both.
- Datasets.