Development of a Stemmer for the Greek Language - SAIS
Development of a Stemmer for the Greek Language - SAIS
Development of a Stemmer for the Greek Language - SAIS
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
We can say that <strong>the</strong> algorithm is working with a quite good precision as <strong>the</strong> error<br />
percentages (7,3% and 8,5%) are considered as acceptable. Based on <strong>the</strong> Table 7<br />
we can conclude that even if we follow an inflection removal technique, without<br />
removing <strong>the</strong> derivational part <strong>of</strong> <strong>the</strong> words, we still have overstemming errors.<br />
These errors occur because <strong>of</strong> <strong>the</strong> large exception words in <strong>the</strong> <strong>Greek</strong> language as<br />
well as <strong>the</strong> high inflectional character <strong>of</strong> <strong>the</strong> language.<br />
In similar experiments on error distribution in Kalamboukis & Nikolaidis (1995)<br />
research, <strong>the</strong>y had an average <strong>of</strong> 17,8% overstemming and 69,7% understemming<br />
(<strong>the</strong> rest 12,5% referred as o<strong>the</strong>r errors). We mention that <strong>the</strong> average <strong>of</strong> <strong>the</strong><br />
precision in that algorithm was 89,6%.<br />
34