CloniZER Spell Checker Adaptive, Language Independent Spell Checker
Loghman Barari, Behrang QasemiZadeh, Digital Clone · 2005
Abstract CloniZER spell checker is an adaptive, language independent and 'built-in error pattern free' spell checker tool which is based on 'Ternary Search Tree' data structure. It suggests the proper form of the misspelled words using nondeterministic traverse. In other words the problem of spell checking is addressed by traverse a tree with variable weighted edges. The proposed method learns media error pattern and improves its suggestions as time goes by. Instead of using expert knowledge for error pattern modelling, the proposed algorithm learns error pattern by interaction with user. 1 Keywords: Spell Checking, Error Pattern, Learning and Adaptation, Lexicon, Ternary Search Tree (TST). 1. Introduction Nowadays, having general access to Internet as a universal phenomenon, electronic texts development and using the text based query interfaces have made the existence of assisting tools for text manipulation inevitable. For example, about 10 to 12 percent of the queries entered in search engines, have dictation errors [1]. Spell Checker have a vast application zone, such as internet search improvement [2][3], correction of errors caused by OCR