The syllabic segmentation is essential for some linguistic or speech recognition applications. Depending
on the language either ruler-based or statistical approach is being used. For Slovak, the statistical
approach seems to be more suitable.
sylseg-sk implements one of the statistical approaches for the syllabic segmentaton. Each input word is
segmented into the syllables. The several possible segmentations are generated and sorted by the
likelihood. If no input file is specified, the standard input is expected. If an input file is used then
the output is written into the file as well. The filename is input filename with the extension
".syllables".
The input-output code page is ISO 8859-2. To use it with different CP use some CP converter and pipes.
For example to have input and output in UTF-8 use (for interactive use): filtermUTF8-iso2iso2-UTF8sylseg-sk or (for batch processing) iconv-fUTF-8-tISO_8859-2|sylseg-sk|iconv-fISO_8859-2-tUTF-8
The performance of the syllabic segmentation depends on the used statistics. To improve the quality of
the segmentation is possible to train the better system with the sylseg-sk-training tool and replace the
original file located in /usr/share/sylseg_sk/sylseg-sk.stats
The design of the sylseg-sk is language independent. With retrained statistics it theoreticaly should
work for any language.