Spelling correction in agglutinative languages

Abstract

This paper presents an approach to spelling correction in agglutinative languages that is based on two-level morphology and a dynamic programming based search algorithm. Spelling correction in agglutinative languages is signi cantly di erent than in languages like English. The concept of a word in such languages is much wider that the entries found in a dictionary, owing to productive word formation by derivational and in ectional a xations. After an overview of certain issues and relevant mathematical preliminaries, we formally present the problem and our solution. We then present results from our experiments with spelling correction in Turkish, a Ural{Altaic agglutinative language.Our results indicate that we can nd the intended correct word in 95% of the cases and o er it as the rst candidate in 74% of the cases, when the edit distance is 1.

Spell Checking and Error Correcting Application for Turkish

“…This study differs from the other studies that have reviewed candidate words for all misspelled words in order to correct misspelled words in Turkish [28]- [31], [34] taking into account the length (n) of analyzed words n, (n-1), and (n-2) grams were used to correct the word. This makes the application optimized for agglutinative languages as if they have a corpus with direct a cyclic word graph tree that has inputs as root word, suffixes and related special cases.…”

“… bigram of "bilgisayar" ; bi-il-lg-gi-is-sa-ay-ya-ar  trigrams of "bilgisayar" ; bil-ilg-gis-isa-say-aya-yar When the studies that perform error checking in Turkish using n-gram and edit distance were analyzed, the unigram, bigram, and trigram types of n-grams were found to be used as text attributes [28]- [32], [34]. In their study [30], error checking is studied, 40 volunteers inscribed 15 different texts of at least 500 words through computer.…”

“…Certain studies have been revealed that perform error checking by examining the morphological structure of the language [24]- [27] by using ngram and edit distance algorithms [28]- [32] and by using text similarity [33]. Some of these studies have generated candidate words for all misspelled words after spell check, in order to correct misspelled words [28]- [30].…”

“…Certain studies have been revealed that perform error checking by examining the morphological structure of the language [24]- [27] by using ngram and edit distance algorithms [28]- [32] and by using text similarity [33]. Some of these studies have generated candidate words for all misspelled words after spell check, in order to correct misspelled words [28]- [30]. The developed software [28] comprises two-steps, determining all the roots from the dictionary that can be the root of the misspelled word, and generating (systematically) all the possible words that "resemble" the given character string, from these roots, realizes the correct word suggestions with a 95% success rate when the count of incorrect spelling letters is one.…”

“…Some of these studies have generated candidate words for all misspelled words after spell check, in order to correct misspelled words [28]- [30]. The developed software [28] comprises two-steps, determining all the roots from the dictionary that can be the root of the misspelled word, and generating (systematically) all the possible words that "resemble" the given character string, from these roots, realizes the correct word suggestions with a 95% success rate when the count of incorrect spelling letters is one. Also it is found that the accuracy of choosing among the corrections, the contextually salient correction as the first option is 74% through a statistical ranking process that takes context into account.…”