<?xml version="1.0" standalone="yes"?>
<Paper uid="C92-1010">
  <Title>PARSING AGGLUTINATIVE WORD STRUCTURES AND ITS APPLICATION TO SPELLING CHECKING FOR TURKISH</Title>
  <Section position="7" start_page="0" end_page="0" type="evalu">
    <SectionTitle>
4. Performance Results
</SectionTitle>
    <Paragraph position="0"> This spelling checker has been i,nl)lenumted in  guage. Its current version takes nearly 600 Kbyl,es including t, he dictionary.</Paragraph>
    <Paragraph position="1"> The checker can be inserted to different word processiagapplieat.ions or can be used separately. We haw' integrated it to GNU-I!EMACS text edit, or for use on IgI'EX document.s, ht this form, the program is avail able for use within the university and around a nun&gt; her of sites oa luternot. It is also I~ossible to obtain solnp statistical hiforniatiou 1) 3&amp;quot; running the progranl willi -s option.</Paragraph>
    <Paragraph position="2"> ()Ill' resilll.s indicate thai the llIlitti)er of distinct words withiil a document is relaliwdy small, and more par ticularly, the percentage of distinct words l.o total words processc(l hicreases as the \[eligill of the docu Inelll decreases. Approxiniately 40% of the ufissl)e\]led word:, are delecled by syllal)ifiealion check and flu' resl ale detected by other checks. The nul'lll)er of disimct words all\]?cl the execulion linie ill()re than Ihc lotai nuuiber of words, as expoclod, because a word is fully aualyzod only ouce If it occtlrs again in Ihe text, the resillt o1&amp;quot; the pt'~,ViOllS check is used, Iu geiiera\]. Ihc slwlling ch,wker can process lit Ill00-::;000 words (roughly 7-6 pages) per s,'eond, depending on Ihe docuutent. The functioilal perforlnance of the.</Paragraph>
    <Paragraph position="3"> spelling checker ('au hi, title trilled \])V analyzing the word \[isl and inserting the additional al)l)ropriale flags</Paragraph>
  </Section>
class="xml-element"></Paper>