Online Konkani spell checker in Devanagari script developed by Dhempe College lecturer

The free to use spell checker has a corpus of 54,295 root words

SHASHWAT GUPTA RAY

[email protected]

- Advertisement -

Add as preferred source on Google

If you enjoyed reading this article, you can help support our journalism by adding OHeraldo as a preferred source.

PANJIM: With growing use of Konkani language in Devanagari script, a senior Computer Science faculty at the Dhempe College has developed an online Konkani spell checker in Devanagari script to help the users deliver error-free work. 

This is the first attempt to develop a spell checker for a low-resource language, Konkani, in Devanagari script with a corpus of 54,295 root words.

“Today we are seeing a lot more publishing work happening in various domains in Konkani language, using Devanagari script. That is why there’s a need to deliver quality work. With this objective, we have developed Konkani spell checker software in Devanagari script to help the users deliver error-free work,” Associate Professor in Computer Science, DCT’s Dhempe 

College of Arts and Science, Miramar Annie Rajan 

- Advertisement -

told O Heraldo.

 Rajan said, “The design and implementation of 

the spell checker was 

challenging.”

- Advertisement -

“Since the analysis is done on the basis of morphology, so the time complexity for giving suggestions to a specific type of wrong words is much faster than a conventional spell checker,” Rajan said.

The spell-checker tool is available online for researchers and other stakeholders. One can visit the website www.konkanispellcheck.com and do spell check.

“In the future, we would like to increase the corpus and f-score for the spell checker,” she said.

“We are also getting responses from users and we are able to further refine the spell checker,” she said.

The spell checker tool has been designed using a Python-based web framework, Django. Versions of Python 3.5 and Django 2.2 have been used. The Konkani dictionary available in XML is parsed using the ‘elementtree’ package and the manual dictionary using the ‘csv’ library. The tool did not rely on any database management system.

“We have made use of two dictionaries where the words are arranged as per their morphology. One is an XML-based dictionary, and the other dictionary has been created manually and contains 2,657 root words. If the inflections are considered then the dictionaries have more than 60,000 words,” the Computer Science faculty said.

The spell checker is at an alpha testing phase, but still, it is made available to the public so that the contribution to the correction of the spell checker can be crowd-sourced. 

“There may be a few words that may appear which may be not correct linguistically in the spell check. We are continuously improving this working tool with the help of the users,” the lecturer said.

Associate Professor in Konkani and Head, Department of Indian Languages, Dhempe College of Arts and Science, Anjul Sakhardande, who has helped in validating the spell checker, said that this tool is a need of the hour.

“For example, it will help those involved in translation work and also PhD scholars who are submitting their thesis work in Konkani. They will be able to do spell check using this tool,” she said.

Share This Article