metadata

license: mit
task_categories:
  - text-classification
language:
  - sw
size_categories:
  - 1K<n<10K
viewer: true

Overview

This dataset contains offensive and non-offensive sentences. The data was scraped from JamiiForums using a prepared wordlist. The dataset contains sentences that consists of swahili abusive words (in the wordlist) but does not contain sarcastic abuse.

Dataset details

The dataset is divided into train, evaluation and test datasets. The training dataset consists of 4954 sentences, evaluation dataset consists of 990 sentences and the test dataset consists of 660 sentences.

Dataset annotations

0: non-offensive
1: offensive