Malayalam News Article Category Classification Dataset

Serving as part of the AI4Bharat-IndicNLP corpus, which contains 2.7 billion words across 10 Indian languages. It is used to create news article category classification datasets for 9 languages, including Malayalam, to evaluate pre-trained word embeddings.

Use Cases

• Train a text classifier on news articles to predict their category labels.
• Evaluate the performance of pre-trained IndicNLP word embeddings on a news classification task.
• Analyze linguistic features and category distributions within the Malayalam news corpus.

Strengths

• Part of a large-scale corpus containing 2.7 billion words.
• Supports 10 Indian languages from two language families.
• Specifically designed for evaluating word embeddings on classification tasks.

Limitations

• The specific number of rows, columns, and sample data for the Malayalam subset is unknown.
• The dataset's creation methodology and potential label sources are not detailed.
• Geographic and temporal coverage for the news articles is unspecified.

Provenance

Source

AI4Bharat-IndicNLP project, author rajeshradhakrishnan.

Collection Method

Part of an ongoing effort to create large-scale, general-domain corpora for Indian languages; news articles were used to create classification datasets.

Freshness

Last updated on 2024-09-10.

Geography

India, focusing on Indian languages including Malayalam.