ABSTRACT

Spelling mistake occurs when a particular word is not spelt according to lexicon of the system where the word is defined. There are application programs called spell checkers used to identify the misspelled words in a document. The prime and desired output of a spell checker program is to suggest correct list of spellings for the misspelled word. From the literature, it has been observed that trend of spelling error analysis, detection of error and correction still area of research for Sindhi language. So far, various word processing applications based on Sindhi language have been developed, but without the functionality of Sindhi spell checking. One of suitable application for Sindhi word processing is Unicode based Sindhi word processor, ln page is also well known word processing application based on Sindhi and Urdu languages. But, these applications are lacking the functionality of spell checker. MS Word is the world's most used software for word processing and also support Sindhi language but lacking the functionality of Sindhi spell checker. The main focus of this research is to build a tool which will check spellings and suggest other relevant words to mistype d Sindhi word. After this research, Microsoft Word add-in will be developed which can be installed in Microsoft Word and spell checking feature will be used. Character n-grams algorithm is used to identify spelling error and creates suggestion list, when error is identified then Edit Distance algorithm is used to replace incorrect word into suggested word. This tool will suggest relevant Sindhi words and correct mistyped words easily just in one click. It scans all Sindhi words in document and scans one by one for correction. Proposed system's accuracy is 82% and average suggestion words is 6 for incorrect word. Proposed system is more efficient than other two spell checkers 'Sindhi Spell Checker for OpenOffice' and 'Urdu Mehfil's' spell checker. This tool will help to check spelling for Sindhi language in MS Word and it also helps to Sindhi newspaper (Kawaish, Koshish, and Awami Awaz), organizations and Sindhi publishers (Sindh Text Book Board).