A Spell-checker Integrated Machine Learning Based Solution for Speech to Text Conversion

Hasan, H. M Mahmudul; Islam, Md. Adnanul; Hasan, Md. Toufique; Hasan, Md. Araf; Rumman, Syeda Ibnat; Shakib, Md. Najmus

doi:10.1109/icssit48917.2020.9214205

Cited by 10 publications

(8 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They used 50,000 ngram sentences to detect and correct the errors. Hasan et al [20] proposed a solution for Bangla Speech to Text(STT) conversion approach in 2020. They also integrated a spell corrector to improve accuracy for the proposed STT system.…”

Section: B Spell Checkermentioning

confidence: 99%

“…• Limited error detection and suggestions by spell checker: The previous spell checkers including the recent ones ( [20] and [21]) can detect only typographic errors. On the contrary, [22] can detect only real-word errors.…”

Section: Limitations Of Previous Approachesmentioning

confidence: 99%

“…Our spell checker uses a large lexicon extracted from our corpus that helps to detect diversified spelling mistakes and to present accurate suggestions for erroneous words. Another flaw in the previous systems except [20] is that they tested the systems using test datasets consisting of single words instead of complete sentences that might not reflect real-life scenarios.…”

Section: Limitations Of Previous Approachesmentioning

confidence: 99%

“…However, they tested their spell checker with datasets containing 1,607 and 2,450 words, respectively. Most of these approaches, including recent work [20] and [21] handle specifically typographic errors, [22] handles particularly real-word errors where the proposed approach handles all different types of errors that exist in Bangla language. Another significant difference between our proposed spell checker and [13] to [22] (excepts [20]) is that they tested their spell checker on test sets consisting of only single words instead of complete sentences, which might not reflect the reallife situations and complexities.…”

Section: Comparison With the Existing Solutionsmentioning

confidence: 99%

“…Most of these approaches, including recent work [20] and [21] handle specifically typographic errors, [22] handles particularly real-word errors where the proposed approach handles all different types of errors that exist in Bangla language. Another significant difference between our proposed spell checker and [13] to [22] (excepts [20]) is that they tested their spell checker on test sets consisting of only single words instead of complete sentences, which might not reflect the reallife situations and complexities. Finally, all the previous approaches, including the recent recurrent neural network (RNN)-based spell checker (82.86%) [22] show In recent days, Bangla Academy changes spellings of many frequently used Bangla words.…”

Section: Comparison With the Existing Solutionsmentioning

confidence: 99%

See 4 more Smart Citations

Development of Bangla Spell and Grammar Checkers: Resource Creation and Evaluation

2021

View full text Add to dashboard Cite

A spell and grammar checker is profoundly essential for diverse publications especially for Bangla language in particular as it is spoken by millions of native speakers around the world. Considering the lack of research efforts, we demonstrate the development of a comprehensive Bangla spell and grammar checker with necessary resources. At first, a full-fledged and generalised Bangla monolingual corpus comprising over 100 million words has been built by scraping reputed, diversified online sources and then an extensive Bangla lexicon consisting of over 1 million unique words has been extracted from that corpus. Based on these corpus and lexicon, we have developed a combined spell and grammar checker application that simultaneously detects distinct spelling and grammatical mistakes and provides appropriate suggestions for both as well. The spell checker uses the Double Metaphone algorithm and Edit distance based on the distributed lexicons and numerical suffix dataset to detect all types of Bangla spelling mistakes with an accuracy rate of 97.21% individually. The grammar checker detects errors based on language model probability i.e. combination of bigram and trigram, and generates suggestions based on the Cosine similarity measure with the accuracy rate of 94.29% individually. The datasets and codes used in this work are freely available at https://git.io/JzJ4w.

show abstract

Section: B Spell Checkermentioning

confidence: 99%