RACAI GEC – A hybrid approach to Grammatical Error Correction

Boroş, Tiberiu; Dumitrescu, Ştefan Daniel; Zafiu, Adrian; Mititelu, Verginica Barbu; Văduva, I.

doi:10.3115/v1/w14-1705

Cited by 2 publications

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, the probability of a learner n-gram is compared with the probability of a candidate corrected ngram, and if the difference is greater than some threshold, an error was perceived to have been detected and a higher scoring replacement n-gram could be suggested. Some teams used this approach only to detect errors, e.g., IPN (Hernandez and Calvo, 2014), which could then be corrected by other methods, whilst other teams used other methods to detect errors first, and then made corrections based on the alternative highest n-gram probability score, e.g., RAC (Boroş et al, 2014). No single team used a uniquely LM-based solution and the LM approach was always a component in a hybrid system.…”

Section: Approachesmentioning

confidence: 99%

The CoNLL-2014 Shared Task on Grammatical Error Correction

Briscoe

et al. 2014

Proceedings of the Eighteenth Conference on Computational Natural Language Learning: Shared Task

395

180

View full text Add to dashboard Cite

The CoNLL-2014 shared task was devoted to grammatical error correction of all error types. In this paper, we give the task definition, present the data sets, and describe the evaluation metric and scorer used in the shared task. We also give an overview of the various approaches adopted by the participating teams, and present the evaluation results. Compared to the CoNLL-2013 shared task, we have introduced the following changes in CoNLL-2014: (1) A participating system is expected to detect and correct grammatical errors of all types, instead of just the five error types in CoNLL-2013; (2) The evaluation metric was changed from F 1 to F 0.5 , to emphasize precision over recall; and (3) We have two human annotators who independently annotated the test essays, compared to just one human annotator in CoNLL-2013.

show abstract

Section: Approachesmentioning

confidence: 99%

The CoNLL-2014 Shared Task on Grammatical Error Correction

Briscoe

et al. 2014

Proceedings of the Eighteenth Conference on Computational Natural Language Learning: Shared Task

395

180

View full text Add to dashboard Cite

show abstract

“…outperform traditional approaches based on edit distance and so on [5] ; Tiberiu Boros proposed RACAI hybrid grammatical error correction system [7] , this system was validated during the participation into the CONLL'14 Shared Task on Grammatical Error Correction. Compared with the traditional method this system overcomes some shortcomings of them, such as can reduce the number of statistical errors and rules; these two methods have made great progress in text semantic correction.…”

mentioning

confidence: 99%

Research and realization of Chinese text semantic correction Based on Rule

Wu¹,

Zhuang²,

Jiang³

et al. 2016

Proceedings of the 2015 3rd International Conference on Education, Management, Arts, Economics and Social Science

View full text Add to dashboard Cite

Automatic correction is an important research field in natural language processing, it is still relatively weak in the text automatic correction technology on semantic level. This paper presents develop XML rules based on LanguageTool Chinese grammar correction, match error correction in semantic level for the content of the rule to write the corresponding rule base, and realize automatic detect the corresponding content of semantic error in the text, and put forward the corresponding modification suggestions and opinions. The experiments show a high correct rate that the process of correction the corpus, which shows that the research and implementation of semantic of Chinese text is very practical significance and valuable. The introduction Background and significance of research. Text is an important carrier of human social information. With the rapid development of whole society information process, the importance and urgency of text information correction is more and more obvious. The former researchers in the text of the technology has made great achievements, but their check wrong technology just based on the word level, dealing with more words, less word or wrong character correction in the text. If testing an essay about the text of population statistics in China, the correct expression is "我国人口统计(不包括台湾、香港、澳门)"("the population statistics of our country (excluding Taiwan, Hongkong, Macao") but appear the "大陆人口统计"("the population statistics of mainland") in the text. If do not found this mistake and report it out, which has violated China's political class information what appeared semantic collocation error. It is difficult to estimate the speed of network transmission, which will bring many negative effects to the society. Semantic errors, text automatic error-detection by word level is unable to achieve the error correction on semantic level, so the research and implementation of Chinese text semantic correction is a key and an essential part of Chinese text automatic correction development. The research status at foreign. In the early 1960s, foreign carried out the study of English text automatic correction; development to today, its technology is very mature already. Because of the English text use the spaces between words and words for separators, their automatic correction is the core of the word, word-error can be of two types, one is non-word error, the other is real-word error [1]. The study found that the non-word errors in the English text accounted for 60%, the real-word error accounted for 40% [2]. The non-word errors that the string is not exist in dictionary [3] ; Real-word error that string is exist in the dictionary [3] , but it with the context collocation error, which cause syntactic semantic error, so the real-word error is the semantic error. Traditional real-word error detection method mainly has two types: based on the traditional method of natural language processing and the method based on statistical language model [4]. It has been studying and improving the technology of semanti...

show abstract

RACAI GEC – A hybrid approach to Grammatical Error Correction

Cited by 2 publications

References 6 publications

The CoNLL-2014 Shared Task on Grammatical Error Correction

The CoNLL-2014 Shared Task on Grammatical Error Correction

Research and realization of Chinese text semantic correction Based on Rule

Contact Info

Product

Resources

About