Automatic Error Detection and Correction in Malayalam

Author(s):

Ambili T , Mar Athanasius College of Engineering, Kothamangalam; Panchami K S, Mar Athanasius College of Engineering, Kothamangalam; Neethu Subash, Mar Athanasius College of Engineering, Kothamangalam

Keywords:

Natural Language Processing (NLP), corpus, N-gram. Tokens, Lexicon

Abstract:

Spelling error correction is a Natural Language Processing (NLP) problem, and it has recently become relevant because many of the potential NLP applications such as text summarization, sentiment analysis and machine translation etc take advantage of spelling error analysis. Spell checking is a well-known task in NLP. Spelling error detection and correction is the process that will check the spelling of words in a document, and in occurrence of any error, list out the correct spelling in the form of suggestions. The proposed method develops a system for spelling error detection and correction in Malayalam. The proposed system uses a spell checker that detects the error by a dictionary lookup approach and error correction is done through N-gram based technique. In dictionary method, it checks each word of input for its presence in the dictionary. If the word is present in the dictionary then it is a correct word else put into the list of error words. N-gram based technique corrects error by finding similarity between words and computing a similarity coefficient. Due to morphological richness of Malayalam, error detection and correction is a challenging task, however the proposed system meets this challenge and has high accuracy compared to other existing approaches.

Other Details:

Manuscript Id	:	IJSTEV3I2002
Published in	:	Volume : 3, Issue : 2
Publication Date	:	01/09/2016
Page(s)	:	92-96

Download Article