README.md

metadata

annotations_creators:
  - self-annotated
language_creators:
  - found
language:
  - tw
license:
  - cc0-1.0
multilinguality:
  - monolingual
size_categories:
  - 10K<n<100K
source_datasets:
  - original
task_categories:
  - translation
task_ids:
  - language-identification
pretty_name: Twi Words
dataset_info:
  features:
    - name: word
      dtype: string
    - name: frequency
      dtype: int32
  splits:
    - name: train
      num_bytes: ...
      num_examples: ...
  download_size: ...
  dataset_size: ...

Twi Words Dataset

The Twi Words Dataset is a curated list of over 50,000 unique words in Twi, a major language spoken in Ghana. This dataset aims to support Natural Language Processing (NLP) tasks in Twi, particularly for low-resource language modeling, classification, and lexicon development.

Purpose

This dataset is created to support:

• NLP development for low-resource African languages
• Spell checkers and autocorrect models
• Text-to-speech and speech-to-text training
• Language modeling and word embeddings
• Educational and lexicographic tools

Data Fields

Each row in the dataset contains:

Word	Frequency
twi_word_1	1234
twi_word_2	987
...	...

• Word: A unique word in Twi extracted from a large corpus.
• Frequency: The number of times the word appears in the corpus, indicating its usage frequency.