| Journal of Information and Communications Technology:
Algorithms, Systems and Applications
Received: 10 August 2025; Revised: 27 December 2025; Accepted: 27 December 2025; Published Online: 29 December 2025.
J. Inf. Commun. Technol. Algorithms Syst. Appl., 2025, 1(3), 25316 | Volume 1 Issue 3 (December 2025) | DOI: https://doi.org/10.64189/ict.25316
© The Author(s) 2025
This article is licensed under Creative Commons Attribution NonCommercial 4.0 International (CC-BY-NC 4.0)
Spell Checker for Low-resource Konkani Language
Annie Rajan,
1,*
Nehal Kalita
2
and Ambuja Salgaonkar
3
1
Department of Computer Science, DCT’s Dhempe College of Arts and Science, Panaji, Goa, 403001, India
2
Independent Researcher, Navi Mumbai, Maharashtra, India
3
Department of Computer Science, University of Mumbai, Mumbai, Maharashtra, 400098, India
*Email: ann_raj_2000@yahoo.com (Annie Rajan)
Abstract
A spell checker is an application that identifies misspelled words by analyzing the sequence of characters in
each word. Spell checking applications exist for many of the Indian languages in the Eighth Schedule of the
Indian Constitution. However, there are not as many spell checkers for languages that were added later, some of
which are low-resource languages. Konkani is one such language. This is the first time a spell checker has been
developed for Konkani, in Devanagari script. Konkani is a macrolanguage, and developing a spell checker is
challenging. We have presented the design and implementation of the spell checker. The proposed approach
makes use of dictionary lookup to identify correct words and minimum edit distance to suggest correct words
for misspelled words. This spell checker also achieved a high F-score after being tested with a set of Konkani
words. It has 1,510,514 unique words in the dictionary.
Keywords: Natural language processing; Indian language; Diacritics; Minimum edit distance; Python.
1. Introduction
India is a country with 22 languages in its Eighth Schedule language list
[1]
to the constitution and 38 languages
that are not on the list. A developing country like India needs the digital footprint of these languages as various
Natural Language Processing (NLP) tools. Building NLP tools like Part-of-Speech (PoS) tagger,
[2]
Named Entity
Recognizer (NER),
[3]
morphological analyzer,
[4]
etc. is a challenging task since to build these tools there is a need
for an annotated corpus in the language in which the tool is built. Information processing in low-resource
languages involves technologies and methods to understand corpora and linguistic databases. By processing
this information, the resources of low-resource languages can be preserved, helping to bridge communication
gaps.
A spell checker is an application that flags words in a document that may not be spelled correctly.
[5,6]
A spell
checker is a basic need of a word processor in any language. The spell checker analyzes the written text in order
to identify any misspellings and gives the best correct suggestions for those misspellings. Spell checking
applications present valid 
document. The user then selects from a list of suggestions or chooses to ignore the suggestions and accept the
current word as valid. Regardless of how often this is done, the spell-checking application will perform its task
independent of the types of misspelled words most commonly made by the user. The spell checkers are crucial
in making quality content without any mistakes or ambiguity. A misspelled word can change the meaning, focus,
and intention of a word and therefore its content, and also can lead to reading and attention discomfort.
The essence of digital applications is exponentially increasing day by day. It is difficult to imagine a regular day
without search engines, social media, online news, emails, and word processing. Further, there are other NLP
applications like speech-to-text and text-to-speech engines, Optical Character Recognition (OCR) systems,
speech synthesizers, and Machine Translation (MT) systems that are evolving. Spell checkers and correctors
play a crucial role in the development of all these applications, and they are deeply coupled with the NLP
ecosystem. Extensive work is reported for English spelling detection and a limited number of Indian languages,
whereas no work has been reported for Konkani, the state language of Goa, India. Konkani is a low-resource
macrolanguage that has multiple scripts and is classified in the linguistic database ISO 639-3.
[7]
The Konkani
language has 36 consonants and 12 vowels. The methods available for other languages cannot be directly
applied to Konkani. One such example is phonetic based spell checking which cannot be directly implemented
in a generalized Konkani spell checker. This is because the pronunciation differs among different variants of
Konkani, like Mangalorean Konkani differs Goan Konkani. In this paper, we have tried to develop a spell checker
for Konkani with 1,510,514 unique words.
[8]
1.1 Literature survey
Spelling checkers are the basic tools needed for word processing and document preparation. It is a tool that
enables users to check the spellings of the words in a text file, validates them, i.e., checks whether they are right
or wrongly spelled, and, in case the spell checker has doubts about the spelling of the word, suggests possible



Languages like Hindi,
[9-13]
Tamil,
[14
−
18]
Punjabi,
[19-21]
Kashmiri,
[22]
Sanskrit,
[23]
Bengali,
[24-28]
Marathi,
[29,30]
Gujarati,
[31,32]
Sindhi,
[33-35]
Telugu,
[13,36]
Kannada,
[37,38]
Malayalam,
[39-44]
Urdu
[45-48]
Assamese,
[49-51]
Manipuri,
[52]
Nepali.
[53-55]
Oriya,
[56,57]
and Bodo,
[58]
have papers in spell checker. In Table 1, the methods used in papers on spell checking
for Indian languages have been listed.
Table 1: Methods used in papers of spell checking and suggestion.
Sr.
No.
Language
Methods
Train data
size in
words
Test data
size in
words
Overall
accuracy
1
Hindi
Minimum edit distance, Statistical machine
translation
*
870
83.2%
2
Hindi
Character n-gram, Dictionary lookup
*
*
*
3
Hindi
Dictionary lookup, Minimum edit distance
~117,000
291
*
4
Hindi
Minimum edit distance
*
*
*
5
Hindi
Statistical machine translation, Convolutional
neural network and Gated recurrent unit
108,587
*
85.4%
6
Tamil
Character bi-gram, Minimum edit distance, Word
frequency, Hashing
4,000,000
2,105
98.4%
7
Tamil
Bloom filter, Minimum edit distance, Long short-
term memory
249,056
*
*
8
Tamil
Dictionary lookup, Character bi-gram, Minimum
edit distance
*
*
89.13%
9
Tamil
Minimum edit distance, Distance matrix
*
*
*
10
Tamil
Minimum edit distance, Rule-based, Soundex,
Long short-term memory
*
*
95.67%
11
Punjabi
Lexicon lookup
~150,000
225
95.61%
12
Punjabi
Lexicon lookup
*
*
83.5%
13
Punjabi
Dictionary lookup, Minimum edit distance
~1,000,000
*
87.2%
14
Kashmiri
Dictionary lookup, Hashing, Binary search tree
~1,000,000
*
~80%
15
Sanskrit
Morphological rules
13,000
1,500
99%
16
Bangla
Partition around medoids clustering
*
2,450
99.8%
17
Bangla
Dictionary lookup, Minimum edit distance
15,162,317
250,000
~95%
18
Bangla
Double metaphone encoding
*
1,607
91.67%
19
Bangla
Finite state automation
*
*
~70%
20
Bangla
Convolutional neural network, Bidirectional
encoder representations from transformers
model
513,000
52,400
87.8%
21
Marathi
Morphological rules
13,000
10,648
99.57%
22
Marathi
Minimum edit distance, Cosine similarity
algorithm
*
929,663
85.88%;
86.76%
23
Gujarati
Minimum edit distance
192,000
79,024
83.14%
24
Gujarati
Mel frequency cepstral coefficients, Gammatone
frequency cepstral coefficients, Bidirectional
encoder representations from transformers
model
*
*
*
25
Sindhi
Minimum edit distance, SoundEx algorithm,
ShapeEx algorithm
100,000+
*
*
26
Sindhi
Minimum edit distance, SoundEx algorithm,
ShapeEx algorithm
250,000
1,744
*
27
Sindhi
Dictionary lookup
70,576
2,052
*
28
Telugu
Statistical machine translation, Convolutional
neural network and Gated recurrent unit
92,716
*
89.3%
29
Telugu
Markov Models
3,300,000+
*
*
30
Kannada
Morphological rules, Dictionary lookup
*
112,000
97.83%
31
Kannada
Morphological rules, Dictionary lookup
3,000,000
43,209
90% noun;
80% verb
32
Malayalam
Long short-term memory
1,505,279
500
55.2%
33
Malayalam
Character n-gram, Minimum edit distance
~80,000
10,000
91%
Sr.
No.
Language
Methods
Train data
size in
words
Test data
size in
words
Overall
accuracy
34
Malayalam
Finite state machine
*
*
*
35
Malayalam
Dictionary lookup, Character n-gram
~10,000
10,000
*
36
Malayalam
Sequence-2-sequence, Lexicon lookup, Hashing,
Character n-gram
10,000
6,000
91.26%
37
Malayalam
Word length difference, Minimum edit distance,
Character n-gram, Phoneme similarity, Random
forest classifier
*
*
71%
38
Urdu
SoundEx algorithm, Single edit distance
1,700,000
724
96.27%
39
Urdu
SoundEx algorithm, ShapeEx algorithm
1,700,000
280
93.5%
40
Urdu
Reverse edit distance
110,582
*
*
41
Urdu
Character n-gram, Minimum edit distance
593,738
510,547
83.67%
42
Assamese
Dictionary lookup, SoundEx algorithm, Hashing,
Minimum edit distance
*
5,000+
*
43
Assamese
Minimum edit distance, Morphological rules,
Dictionary lookup
16,000
1,000;
5,000;
10,000
0.58; 0.63;
0.69 recall
44
Assamese
Character tri-gram
220,743
500
76.6%
45
Manipuri
Dictionary lookup, Reverse dictionary lookup,
Minimum edit distance, Phonetic encoding
~10,000
*
*
46
Nepali
Character bi-gram, Bidirectional encoder
representations from transformers model,
Probabilistic spelling correction
~350,000
~125,000
69.1%
47
Nepali
Dictionary lookup, Minimum edit distance,
Decision Tree
*
*
78%
48
Nepali
Gated recurrent unit
120,000
*
73%
49
Odia
Confusion matrix, Character n-gram
*
685
88%
50
Odia
Dictionary lookup, Minimum edit distance
36,000
*
*
51
Bodo
Morphological rules, Dictionary lookup
~15,000
~1,500,000
*
* Values not listed
Currently, there are no papers available on spell checking for the Konkani language. The proposed spell-
checking approach discussed in this study is mainly based on dictionary lookups and minimum edit distance.
[59]
2. Methodology
The Konkani language has 15 diacritics.
[60]
A diacritic is a mark added to a letter to show a change in
pronunciation, tone, or emphasis. In Konkani, diacritics in the Devanagari script help distinguish sounds. For
example, in the word (k
), the nasal diacritic  (anusvara) above (
it apart from (
is set to 0.5, and the count of other alphabets is set to 1. Fig. 1 shows the workflow of the proposed approach.
Fig. 1: Flowchart of the Konkani spell checker.
2.1 Dataset
The data used to generate our dictionary originated from the author of Ref. 3, supplemented by a collection of
~4,000 additional words, totaling 1,510,514 words.
2.2 Dictionary creation
The dictionary developed for the spell checker has the words arranged in sequential order based on their first
character, words having the same first character are further arranged in ascending order of length. Three comma
separated valued files were created and were used as reference for executing the proposed approach.
The first file, File-1, is the dictionary, where each row consists of the Konkani words and their respective
character count in the aforementioned order. A sample demonstration is shown in Table 2. In the second file,
File-2, each row consists of unique alphabets, the beginning and ending row numbers in File-1 starting with the
current alphabet, and the minimum and maximum character count of the words starting with the alphabet. A
sample demonstration is shown in Table 3. The third file, File-3, contains the range of rows in File-1 covered by
words beginning with each alphabet, where the range is further subdivided as per character count. A sample
demonstration is shown in Table 4
rows in File-1, that has words with the specified character count. The value -1 for the Konkani alphabet,
means that there are no words with lengths 4.0, 4,5 and 5.0 that begins with this alphabet.
Table 2: Sample rows from File-1.
Row no.
Word
Character count
0

2.0
1

2.0
295189
1.5
334068
  
8.0
334069

1.5
Table 3: Sample range of rows in File-2.
Alphabet
Start row no.
End row no.
Min. character count
Max. character count
0
60449
2.0
8.0
295189
334068
1.5
8.0
Table 4: Sample range of rows in File-3.
Alphabet
4.0 start row
no.
4.0 end row no.
4.5 start row
no.
4.5 end row no.
5.0 start row
no.
5.0 end row no.
189436
191986
191987
197007
197008
204162
296000
297169
297170
299281
299282
302316
-1
-1
-1
-1
-1
-1
Although File-2 contains less information than File-3, this file reduces computation time to identify whether an
input word is incorrect based on its stored list of character count range for each alphabet. If an input word has
character count beyond the specified range in File-2, then it is declared incorrect. Whereas if the character count
is within the range, then File-3 is referred for further analysis of correctness.
2.3 Pipeline of code
Table 5 shows the various abbreviations used while explaining the pipeline of code. The workings of our code
are briefly explained with the help of the five steps mentioned below.
(1) Check if w
i
can be traced in File-1 by iterating through the range of words having the same α as w
i
. If it is
traceable, then further steps should not be continued.
(2) Identify w
s_l
based on w
i_l
.
If w
i_l
is 1.5 or 2.0, then w
s_l
can be within range → (1.0, 2.0); if w
i_l
is 3.0, then w
s_l
can be within range → (2.0,
4.0); if w
i_l
is 3.0, then w
s_l
can be within range → (2.0, 4.0); if w
i_l
is 3.5, then w
s_l
can be within range → (2.5,
4.5); if w
i_l
is 4.0, then w
s_l
can be within range → (3.0, 5.0); if w
i_l
is > 4.0, then w
s_l
can be within range →
( ((ceiling value of w
i_l
) * 0.8), (w
i_l
/ 0.8) ).
(3) Compare w
i
with words from File-1 whose length is within the range of w
s_l
.
If w
i_l
is > 4.0, then s
p
is 80; if w
i_l
is 4.0, then s
p
is 75; if w
i_l
is 3.0, then s
p
is 66.66; if w
i_l
is 1.5 or 2.0, then s
p
is
50 (if s
c
is 1.0) or 75 (if s
c
> 1.0).
(4) Sort w
s
as per s
p
in descending order.
(5) Display the sorted w
s
.
Table 5: List of abbreviations.
Sr. No.
Abbreviations
Full form
1
w
i
input word
2
α
first alphabet of a word
3
w
i_l
length of input word
4
w
s
suggested words
5
w
s_l
expected length of suggested words
6
w
d
word from dictionary as per w
s_l
7
s
p
final minimum similarity value of w
i
and w
s
in percentage
8
s
c
initial count of similar characters between w
i
and w
d
in
sequence
9
w
ln
the longer word among w
i
and w
d
10
w
st
the shorter word among w
i
and w
d
If the list of w
s
contains s
p
values > 80, then display the first ten w
s
as most preferred suggestions and the
remaining as other probable suggestions.
In step (3), the calculation of similarity in terms of the comparison is done in two steps.
The first step is to determine s
c
. There are two methods, and these are demonstrated below with the help of
pseudocode.
Method A:
Method B:
In method A, the counter for w
st
is not updated if the current character of w
st
is not equal to the current character
of w
ln


whereas the counter for the co

In method B, the counters for both w
st
and w
ln
are incremented in every iteration, regardless of the equality of

       do not match, but the succeeding
s
c1
= 0 // s
c
value by method A
c
ln
= 0 // counter for w
ln
c
st
= 0 // counter for w
st
while (c
ln
< length of w
ln
) {
if (c
st
< length of w
st
) and (w
st
[c
st
] == w
ln
[c
ln
])
{
update s
c1
as per conditions
increment c
st
by 1
increment c
ln
by 1
}
else {
increment c
ln
by 1
}
}
s
c2
= 0 // s
c
value by method B
c
ln
= 0 // counter for w
ln
c
st
= 0 // counter for w
st
while (c
ln
< length of w
ln
) {
if (c
st
< length of w
st
) and (w
st
[c
st
] == w
ln
[c
ln
])
{
update s
c2
as per conditions
}
increment c
st
by 1
increment c
ln
by 1
}

lesser value of s
c
as only the first six characters would be identified as similar. So, the larger value among s
c1
and
s
c2
is considered the value of s
c
.
The second step of the calculation of similarity is to determine s
p
. For this, the formula (s
c
/ length of w
ln
) * 100
is used, and the value obtained is compared with different conditions of percentage values depending on w
i_l
. If
conditions are satisfied, then that value is considered the value of s
p
.
2.4 Spell checker tool
The Konkani Spell Check (https://konkanispellcheck.pythonanywhere.com/) tool has been designed using a
Python (version 3.5) based web framework, Django (version 2.2). The Konkani dictionary is parsed using the
         tabase management system. Its user interface incorporates
responsive web design, so it changes based on the screen resolution. Fig. 2 shows the output of words getting
-input
text box and tries to suggest some correct alternatives in the suggested-words text box.
Fig. 2: Analysis through Konkani spell check tool.
3. Results and discussion
A test dataset of 4,853 unique Konkani words from books of Standards 1, 2, and 3 by SCERT-Goa
[61]
was
analysed, out of which 4,041 words were identified as correct by our dictionary. Out of 812 unidentified words,
522 words were assigned suggestions by our system. The F-score value in Table 7 is calculated as per the
confusion matrix outcomes shown in Table 6. The analysis of this dataset took 4,600.15 seconds on a local
machine with an Intel Core i7-3770 CPU and 1600 MHz RAM. The analysis of only the 4,041 identified words
was significantly faster, taking 605.77 seconds. The algorithm analyzes correct words faster because it doesn't
need to perform word identification for suggestion.
Table 6: Confusion matrix.
Outcome
Value
True Positive (TP)
4,041
False Positive (FP)
0
True Negative (TN)
0
False Negative (FN)
812
Table 7: Classification metrics.
Metric
Formula
Value
Precision (P)
TP / (TP + FP)
1
Recall (R)
TP / (TP + FN)
0.833
F-score
2 * (P * R) / (P + R)
0.909
Table 8 presents a comparison of different incorrect word inputs for a proper noun,  (g󰕁󰡂󰕁p󰕁󰕁). Since
this noun is of length 4.0 units, the expected s
p
is 75%. Hence, in Fig. 2, it is shown as a suggestion for 1
st
and 2
nd
input words by the application.
Table 8: Comparison of incorrect inputs.
Sr. no.
Input
Length
Similarity (in %)
1
  (g󰕁󰡂󰠭󰥖󰕁󰕁)
4.5
88.88
2
 (g󰕁󰡂󰕁p󰕁󰡂󰕁)
4.0
75
3
 (g󰕁󰡂󰕁p󰕁󰕁󰠭󰥖
5.5
72.72
4
  (g󰕁󰠭󰥖󰕁󰕁
4.5
66.66
4. Conclusion
The results demonstrate that our spell checker achieved an F-score of 0.909 on a dataset consisting of words
from school books of Standards 1, 2 and 3. This makes our spell checker a reliable source for checking basic
Konkani words. The implementation of an alphabet and word length-based arrangement of the dictionary
makes this spell checker much faster in giving suggestions to wrong words than a conventional spell checker
where no such arrangements are implemented in its dictionary. Future experiments can emphasize on
implementing a much faster algorithm for correct word suggestions and increasing the size of the corpus. A
bigger corpus can help a spell checker reliable in checking advanced Konkani words, the ones that are used in
school books of higher standards.
Funding Declaration
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit
sectors.
Data Availability Statement
The test data is available in a Harvard Dataverse repository (https://doi.org/10.7910/DVN/ECWMBB).


Artificial Intelligence (AI) Use Disclosure
The authors declare that artificial intelligence (AI)-assisted tools were used only for language refinement,
grammar improvement, and manuscript structuring purposes during the preparation of this work. All technical
content, experimental implementation, results, and interpretations were independently developed and verified
by the authors.


References
[1]
Ministry of Home Affairs Government of India, 2017, accessed 08 August 2024,
https://www.mha.gov.in/sites/default/files/EighthSchedule_19052017.pdf,
[2]
H. Li, H. Mao, J. Wang, Part-of-speech tagging with rule-based data preprocessing and transformer,
Electronics, 2022, 11, 56, doi: 10.3390/electronics11010056.
[3]
W. L. Seow, I. Chaturvedi, A. Hogarth, R. Mao, E. Cambria, A review of named entity recognition: from
learning methods to modelling paradigms and tasks, Artificial Intelligence Review, 2025, 58, 315
https://doi.org/10.1007/s10462-025-11321-8
[4]
E. B. Boltayevich, H. S. Mirdjonovna, A. X. Ilxomovna, Methods for Creating a Morphological Analyzer.
In: Zaynidinov, H., Singh, M., Tiwary, U.S., Singh, D. (eds) Intelligent Human Computer Interaction. IHCI
2022. Lecture Notes in Computer Science, vol 13741. Springer, Cham, doi: 10.1007/978-3-031-27199-
1_4.
[5]
A. Toleu, G. Tolegen, R. Mussabayev, A. Krassovitskiy, I. Ualiyeva, Data-driven approach for
spellchecking and autocorrection, Symmetry, 2022, 14, 2261, doi: 10.3390/sym14112261
[6]
N. Hossain, S. Islam, M. N. Huda, Development of bangla spell and grammar checkers: resource creation
and evaluation, IEEE Access, 2021, 9, 141079-141097, doi: 10.1109/ACCESS.2021.3119627.
[7]
ISO 639-2, Library of Congress, 2017, https://www.loc.gov/standards/iso639-2/php/code_list.php,
accessed 10 April 2024
[8]
S. N. Desai, Design and implementation of algorithms for morphology learning and its applications,
Doctoral dissertation, Goa University, 2017.
[9]
B. Kaur, H. Singh, Design and implementation of HINSPELL- Hindi spell checker using hybrid approach,
International Journal of Scientific Research and Management, 2015, 3, 20582061.
[10]
A. Jain, M. Jain, Detection and correction of non-word spelling errors in Hindi language, 2014
International Conference on Data Mining and Intelligent Computing (ICDMIC), Delhi, India, 2014, 1-5,
doi: 10.1109/ICDMIC.2014.6954235.
[11]
A. Sharma, P. Jain, A. Mukerjee: Hindi spell checker Project, Indian Institute of Technology Kanpur,
2013.
[12]
S. Rachel, S. Vasudha, T. Shriya, K. Rhutuja, L. Gadhikar, Vyakranly: Hindi grammar & spelling errors
detection and correction system, 2023 5th Biennial International Conference on Nascent Technologies
in Engineering (ICNTE), Navi Mumbai, India, 2023, 1-6, doi: 10.1109/ICNTE56631.2023.10146610.
[13]
P. Etoori, M. Chinnakotla, R. Mamidi, Automatic spelling correction for resource-scarce languages using
deep learning, Proceedings of ACL 2018- Student Research Workshop, ACL, Melbourne 2018, 1, 146
152.
[14]
K. Uthayamoorthy, K. Kanthasamy, T. Senthaalan, K. Sarveswaran, G. Dias, DDSpell-A data driven spell
checker and suggestion generator for the Tamil language, 2019 19th International Conference on
Advances in ICT for Emerging Regions (ICTer), Colombo, Sri Lanka, 2019, 1-6, doi:
10.1109/ICTer48817.2019.9023698.
[15]
S. Murugan, T. A. Bakthavatchalam, M. Sankarasubbu, SymSpell and LSTM based spell-checkers for
Tamil, Tamil Internet Conference, 2020.
[16]
J. Segar, K. Sarveswaran, Contextual spell checking for Tamil language, 14th Tamil Internet Conference,
1-5, 2015.
[17]
P. Kumar, A. Kannan, N. Goel, Design and implementation of NLP-based spell checker for the Tamil
language., 1st International Electronic Conference on Applied Sciences, Noida, 2020, 16.
[18]
A. Sampath, V. Shanmugavel, Hybrid Tamil spell checker with combined character splitting,
Concurrency and Computation Practice and Experience, 2022, 35, e7440, doi: 10.1002/cpe.7440
[19]
G. S. Lehal, Design and implementation of Punjabi spell checker, International Journal of Systemics,
Cybernetics and Informatics, 2007, 3, 7075.
[20]
J. Kaur, K. Garg, Hybrid approach for spell checker and grammar checker for Punjabi, International
Journal of Computer Science and Software Engineering, 2014, 4, 6267.
[21]

and Technology, 2010.
[22]
A. Lawaye, B. S. Purkayastha, Design and implementation of spell checker for Kashmiri, International
Journal of Scientific Research, 2016, 5, 199200.
[23]
N. Tapaswi, Morphological-based spellchecker for Sanskrit sentences, International Journal of Scientific
& Technology Research, 2012, 1, 14.
[24]
P. Mandal, B. M. M. Hossain, Clustering-based Bangla spell checker, 2017 IEEE International Conference
on Imaging, Vision & Pattern Recognition (icIVPR), Dhaka, Bangladesh, 2017, 1-6, doi:
10.1109/ICIVPR.2017.7890878.
[25]
B. B. Chaudhuri, Reversed word dictionary and phonetically similar word grouping based spell-checker
to Bangla text, Proc. LESAL Workshop Mumbai, 2001.
[26]
N. UzZaman, M. Khan, A double metaphone encoding for Bangla and its application in spelling checker,
2005 International Conference on Natural Language Processing and Knowledge Engineering, Wuhan,
China, 2005, 705-710, doi: 10.1109/NLPKE.2005.1598827.
[27]
M. M. Abdullah, M. Z. Islam, M. Khan, Error-tolerant finite-state recognizer and string pattern
similarity-based spelling-checker for Bangla, Proceeding of 5th international conference on natural
language processing, ICON, 2007.
[28]
C. R. Rahman, M. H. Rahman, S. Zakir, M. Rafsan, M. E. Ali, Spell- A CNN-blended BERT based Bangla
spell checker, Proceedings of the First Workshop on Bangla Language Processing, Singapore, 2023, 1,
717.
[29]
V. Dixit, S. Dethe, R. K. Joshi, Design and implementation of a morphology-based spellchecker for
Marathi, and Indian language, Archives of Control Science, 2005, 15, 251258.
[30]
K.T. Patil, R. P. Bhavsar, B. V. Pawar, Contrastive study of minimum edit distance and cosine similarity
measures in the context of word suggestions for misspelled Marathi words, Multimedia Tools and
Applications, 2022, 82, 1557315591, doi: 10.1007/s11042-022-13948-z.
[31]
H. Patel, B. Patel, K. Lad, Jodani- A spell checking and suggesting tool for Gujarati language, 2021 11th
International Conference on Cloud Computing, Data Science & Engineering (Confluence), Noida, India,
2021, 94-99, doi: 10.1109/Confluence51648.2021.9377072.
[32]
M. Dua, B. Bhagat, S. Dua, An amalgamation of integrated features with Deep-Speech2 architecture and
improved spell corrector for improving Gujarati language ASR system, International Journal of Speech
Technology, 2024, 27, 87-99.
[33]
Z. Bhatti, I. A. Ismaili, D. N. Hakro, W. J. Soomro, Phonetic-based Sindhi spellchecker system using a
hybrid model, Digital Scholarship in the Humanities, 2016, 31, 264282, doi: 10.1093/llc/fqv005.
[34]
I. A. Dahar, F. Abbas, U. Rajput, A. Hussain, F. Azhar, An efficient Sindhi spelling checker for microsoft
word. International Journal of Computer Science and Network Security, 2018, 18, 144150.
[35]
M. Umair, M. U. Rahman, Analysis of Sindhi spelling error patterns for spelling error detection and
correction, International Conference on Computer & Emerging Technologies, 2013.
[36]
K. N. Murthy, Technology for Telugu, Bhasha, 2008, 1, 7095
[37]
S. R. Murthy, V. Madi, D. Sachin, P. R. Kumar, A non-word Kannada spell checker using morphological
analyzer and dictionary lookup method, International Journal of Engineering Sciences & Emerging
Technologies, 2012, 2, 4352, doi: 10.18653/v1/P18-3021
[38]
S. R. Murthy, A. N. Akshatha, C. G. Upadhyaya, P. R. Kumar, Kannada spell checker with sandhi splitter.
2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI),
Udupi, India, 2017, 950-956, doi: 10.1109/ICACCI.2017.8125964.
[39]
S. Sooraj, K. Manjusha, M. A. Kumar, K. P. Soman, Deep learning-based spell checker for Malayalam
language, Journal of Intelligent & Fuzzy Systems, 2028, 34, 1427-1434.
[40]
P. H. Hema, C. Sunitha, Malayalam spell checker using n-gram method, Computational Intelligence in
Data Mining, Advances in Intelligent Systems and Computing, 2016, 1, 217225, doi: 10.1007/978-81-
322-2734-2_23.
[41]
N. Manohar, P. T. Lekshmipriya, V. Jayan, V. K. Bhadran, Spellchecker for Malayalam using finite state
transition models, 2015 IEEE Recent Advances in Intelligent Computational Systems (RAICS),
Trivandrum, India, 2015, 157-161, doi: 10.1109/RAICS.2015.7488406.
[42]
T. Ambili, S. K. Panchamim, N. Subash, Automatic error detection and correction in Malayalam,
International Journal for Science Technology and Engineering, 2016, 3, 9296.
[43]
D. J. Ratman, A. K. Karthika, K. Praveena, R. Tania, S. Thara, N. Prema, Phonogram-based automatic typo
correction in Malayalam social media comments, Procedia Computer Science, 2024, 233, 391400, doi:
10.1016/j.procs.2024.03.229.
[44]
S. Dhanya, M. R. Kaimal, P. Nedungadi, Automatic spelling error classification in Malayalam, ICT-Cyber
Security and Applications, In: Joshi, A., Mahmud, M., Ragel, R.G., Kartik, S. (eds) ICT: Cyber Security and
Applications. ICTCS 2022, Lecture Notes in Networks and Systems, 2024, 916, 301313, Springer
Nature, LNNS 2024, Springer, Singapore, doi: 10.1007/978-981-97-0744-7_25
[45]

& Emerging Sciences, 2004.
[46]
T. Naseem, S. Hussain, A novel approach for ranking spelling error corrections for Urdu, Language
Resources and Evaluation, 2007, 41, 117128, doi: 10.1007/s10579-007-9028-6.
[47]
S. Iqbal, W. Anwar, U. I. Bajwa, Z. Rehman, Urdu spell checking: Reverse edit distance approach,
Proceedings of the 4th workshop on South and Southeast Asian natural language processing, 2013, 4,
5865.
[48]
R. Aziz, M. W. Anwar, M. H. Jamal, U. I. Bajwa, A. K. Castilla, C. U. Rios, E. B. Thompson, I. Ashraf, Real
word spelling error detection and correction for Urdu language, IEEE Access, 2023, 11, 100948
100962, doi: 10.1109/ACCESS.2023.3312730.
[49]
M. Das, S. Borgohain, J. Gogoi, S. B. Nair, Design and implementation of a spell checker for Assamese,
In: Language Engineering Conference, Language Engineering Conference, 2002. Proceedings,
Hyderabad, India, 2002, 156-162, doi: 10.1109/LEC.2002.1182303.
[50]
K. Kashyap, Luitspell- development of an Assamese language spell checker for open office writer,
European Journal of Advances in Engineering and Technology, 2015, 2, 135138.
[51]
R. Choudhury, N. Deb, K. Kashyap, Context-sensitive spelling checker for Assamese language, Recent
developments in machine learning and data analytics, In: J. Kalita, V. Balas, S. Borah, R. Pradhan, R.
(eds) Recent Developments in Machine Learning and Data Analytics. Advances in Intelligent Systems
and Computing, Springer, Singapore, 2019, 740, 177188, doi: 10.1007/978-981-13-1280-9_18.
[52]
H. M. Devi, T. Keat, B. B. Chaudhuri, Spelling Correction in Manipuri text. Advanced Computing
Applications Databases and Networks, Narosa, Silchar, 2011, 2129.
[53]
N. Luitel, N. Bekoju, A. K.Sah, S. Shakya, Contextual spelling correction with language model for low-
resource setting, International Conference on Inventive Computation Technologies, IEEE, Lalitpur,
2024, 7, 582589.
[54]
B. Devkota, B. Adhikar, D. Shrestha, Integrating romanized Nepali spellchecker with SMS based decision
support system for Nepalese farmers, 9th International Conference on Software, Knowledge,
Information Management and Applications, IEEE, Kathmandu, 2016, 3, 16.
[55]
B. Prasain, N. Lamichhane, N. Pandey, P. Adhikari, P. Mudbhari, Nepali spelling checker, Journal of
Engineering and Sciences, 2022, 1, 128130.
[56]
Y. Mohapatra, A. K. Mishra, Spell Checker for OCR, International Journal of Computer Science and
Information Technologies, 2013, 4, 9197.
[57]
A. Pradhan, S. S. Dalai, Design of Odia spell checker with word prediction, International Journal of
Engineering Research and Technology, 2020, 8, 14.
[58]
B. Bhatima, C. S. Prabha, Linguistic foundations for Bodo spell checker, Doctoral dissertation, Gauhati
University (2016).
[59]
E. S. Ristad, P. N. Yianilos, Learning string-edit distance, IEEE Transactions on Pattern Analysis and
Machine Intelligence, 1998, 20, 522532, doi: 10.1109/34.682181.
[60]
Diacritics, 1997, University of Sussex,
https://www.sussex.ac.uk/informatics/punctuation/misc/diacritics, accessed 16 June 2025.
[61]
S.C.E.R.T. Government of Goa, https://scert.goa.gov.in/, accessed 16 June 2025.
Publisher Note: The views, statements, and data in all publications solely belong to the authors and
contributors. GR Scholastic is not responsible for any injury resulting from the ideas, methods, or products
mentioned. GR Scholastic remains neutral regarding jurisdictional claims in published maps and institutional
affiliations.
Open Access
This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which
permits the non-commercial use, sharing, adaptation, distribution and reproduction in any medium or format,
as long as appropriate credit to the original author(s) and the source is given by providing a link to the Creative
Commons License and changes need to be indicated if there are any. The images or other third-party material
in this article are included in the article's Creative Commons License, unless indicated otherwise in a credit line
to the material. If material is not included in the article's Creative Commons License and your intended use is
not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this License, visit: https://creativecommons.org/licenses/by-
nc/4.0/
© The Author(s) 2025