Vol. 10, Issue 1, (issn: 2278-7720)

FiniteCheck
, a grammar checking system for detecting errors in primary school children’s texts written in the Swedish 
language was developed by Hashemi (2003). A finite state approach was used by this system and it had only positive rules 
and no rules described the error structures to be detected. No part of speech tagger for disambiguating POS tagging 
information was used; rather it saved all the possible POS tags for words and disambiguates some of this information in 
parsing phase using filtering transducers. There are components of its grammar; a narrow grammar that accepted only 
grammatical patterns and a broad grammar with relaxed rules were used to parse both grammatical and ungrammatical 
structures. Those sentences or segments that could be parsed using broad grammar but not narrow grammar were marked as 
erroneous. This system found errors related to noun phrase agreement, and use of finite and non-finite verb forms in main 
and subordinate clauses.
Another rule based system for English text was proposed by Rider (2005). In this system, both the manually constructed 
error rules and randomly generated rules were used for error detection in English texts. The manually constructed rules 
worked on the POS tagged text and if a match was found then that particular segment was marked as erroneous. The random 
rules generated were then tested on a corpus of correct English and all those rules that flagged errors in that corpus were 
removed from the set of error rules.

---

www.ijcait.com International Journal of Computer Applications & Information Technology 
Vol. 10, Issue 1, (ISSN: 2278-7720)
P a g e
| 219
A grammar checking system for Brazilian Portuguese language was proposed by Kinoshita et al. (2006). This system was 
developed for use in OpenOffice. Local error rules were applied to outputs of POS tagger and chunker. Structural error rules 
were applied on the outcome of grammatical relation finder (establishes subject, verb, and predicate relations). 
Another rule based grammar checking system for Nepali language was developed by Bal and Shrestha (2007). Various types 
of grammatical errors covered by this system were nominal and verbal agreement, structural errors (for clause and sentence 
structure). This system included a tokenizer, morphological analyzer, POS tagger, chunker/parser, syntax checker etc. 
A rule based grammar checker was developed for Persian language by Ehsan and Faili (2010). Hand crafted rules were 
applied on the tagged input text. Another grammar checking system for Arabic/Persian language was developed by Shaalan 
(2005). In this system, a rule based chart parser was used. Rules were developed to check the agreement of verb with 
particles. These rules were implemented in the form of constraints. Therefore, if a particular constraint is not satisfied, the 
system will generate an error message. 
A rule based grammar checker for Afan Oromo (language widely spoken and used in Ethiopia) was developed by Tesfaye 
(2011). A set of 123 hand crafted rules was constructed. The set contained the rules related to match the grammatical 
agreement between subject and verb, subject and adjective, main verb and subordinate verb in terms of number, gender and 
tense. The system showed an overall precision of 88.89% and a recall of 80%.
A rule based Chinese grammar checker was developed byJiang et al. (2011). A number of hand crafted rules were 
developed. Some of these rules were related with the misuse of quantifier and particle. Some others were used to check the 
mismatch between various word classes like mismatch between verb and object.
A rule based grammar checking system for Malay language was proposed by Kasbon et al. (2011). The Tatbahasa Devan 
corpus was used to obtain the rules of Malay language. Before performing the grammar checking, the system performed two 
additional tasks.
A rule based grammar checker for Punjabi language was developed by Singh and Lehal (2008). An exhaustive set of hand 
crafted rules were created and the input sentences were checked against these rules. These rules were designed to check the 
grammatical agreement between subject and verb, noun and its modifier etc. in terms of number, gender and case. Many 
other components like pre-processor, morphological analyzer, POS tagger, phrase chunker etc. were also developed. They 
used agreement matching techniques for grammar checking.