---
license: cc-by-sa-4.0
language:
- ml
size_categories:
- 10K<n<100K
configs:
- config_name: common_words
  data_files: lexicons/commonwords.tsv
  sep: "\t"
  default: true
- config_name: verbs
  data_files: lexicons/verbs.tsv
  sep: "\t"
- config_name: nouns
  data_files: lexicons/nouns.tsv
  sep: "\t"
- config_name: person_names
  data_files: lexicons/person-names.tsv
  sep: "\t"
- config_name: place_names
  data_files: lexicons/place-names.tsv
  sep: "\t"
- config_name: pronouns
  data_files: lexicons/pronouns.tsv
  sep: "\t"
- config_name: proper_nouns
  data_files: lexicons/proper_nouns.tsv
  sep: "\t"
- config_name: english
  data_files: lexicons/english.tsv
  sep: "\t"
- config_name: english
  data_files: lexicons/english.tsv
  sep: "\t"
- config_name: sanskrit_nouns
  data_files: lexicons/nouns-sanskrit-origin.tsv
  sep: "\t"
task_categories:
- text2text-generation
---
# Malayalam Pronunciation Dictionary

This Dataset has an alternate name of **Malayalam Phonetic Lexicon**. It is curated from the original source [here](https://gitlab.com/kavyamanohar/malayalam-phonetic-lexicon) 

It gives Phonemic transcription of Malayalam words in [IPA](https://en.wikipedia.org/wiki/International_Phonetic_Alphabet) format.

## Dataset Details

### Dataset Description

This is a collection of Malayalam words and their pronunciation described in IPA format. The pronunciations has been automatically generated using [Mlphon]
(https://pypi.org/project/mlphon/) Python library.


- **Curated by:** [Kavya Manohar](https://huggingface.co/kavyamanohar)
- **Language(s) (NLP):** Malayalam
- **License:** CC-BY-SA-4.0

### Dataset Sources [optional]

The Malayalam words in this dataset are categorized into:
  - commonwords
  - verbs
  - nouns
  - english
  - nouns_sanskrit
  - proper_nouns
  - pronouns
  - person-names
  - place-names

The commonwords are ordered by frequency of occurrence in [Indic-NLP-Corpus](https://github.com/AI4Bharat/indicnlp_corpus)
All other categories of words were derived from curated collection of words in [Mlmorph](https://morph.smc.org.in/)

## Uses


### Direct Use

Can be used directly in ASR and TTS systems, where IPA based phonetic lexicon is needed.

Can be used to train Malayalam-IPA transliteration models


## Dataset Structure

Apart from the categorization, no separate train-test splits are provided



## Bias, Risks, and Limitations

- The IPA pronunciation has been derived automatically using Mlphon library. It is not human-curated.

- Mlphon produces [Phonemic transcription, not phonetic](https://verbit.ai/transcription/phonemic-transcription-101-visualizing-sound/). Hence നന്ദി is transliterated as n̪ a n̪ d̪ i  and not  n̪ a n̪ n̪ i. This is the expected behaviour.

- The perfromance of Mlphon on a hand curated gold-standard lexicon has been evaluated and reported in [Mlphon](https://ieeexplore.ieee.org/document/9877808) paper.


![image/png](https://cdn-uploads.huggingface.co/production/uploads/638f810f75d1443c884243b0/CtCaWjw5hmY8DfR4tDl2W.png)

The words in Gold standard lexicon covers 26\% of words in Indic-NLP-Corpus


![image/png](https://cdn-uploads.huggingface.co/production/uploads/638f810f75d1443c884243b0/p8e5dhSOOgLCrTzPoZJYp.png)
Distribution of errors in IPA transcription for different word types in gold standard lexicon.


## Citation


**BibTeX:**

```
@ARTICLE{9877808,
  author={Manohar, Kavya and Jayan, A. R. and Rajan, Rajeev},
  journal={IEEE Access}, 
  title={Mlphon: A Multifunctional Grapheme-Phoneme Conversion Tool Using Finite State Transducers}, 
  year={2022},
  volume={10},
  number={},
  pages={97555-97575},
  keywords={Speech recognition;Knowledge based systems;Vocabulary;Computational modeling;Text processing;Phonetics;Automatic speech recognition;Linguistics;Natural language processing;Computational Phonology;Low Resource Languages;Pronunciation Lexicon;Malayalam;Software Tool;Speech Recognition;Syllabification},
  doi={10.1109/ACCESS.2022.3204403}}
```

**APA:**

K. Manohar, A. R. Jayan and R. Rajan, "Mlphon: A Multifunctional Grapheme-Phoneme Conversion Tool Using Finite State Transducers," in IEEE Access, vol. 10, pp. 97555-97575, 2022, doi: 10.1109/ACCESS.2022.3204403.


## For Further Reading:

- [Blog-1](https://kavyamanohar.com/post/malayalam-phonetic-lexicon)
- [Blog-2](https://kavyamanohar.com/post/malayalam-g2p-mlphon-paper/)
- [Paper]((https://ieeexplore.ieee.org/document/9877808))

## Dataset Card Authors

Kavya Manohar

## Dataset Card Contact

Kavya Manohar