kimrichies/English-Luganda_Bible_corpus

Dataset Card for Dataset Name

This is a parallel dataset for English and Luganda Languages.

Dataset Description

Luganda, also known as Ganda, is a Bantu language primarily spoken in the Buganda region of Uganda. It is one of the main languages of Uganda and serves as an official language alongside English and Swahili. Historically, Luganda was the linguistic and cultural backbone of the powerful Buganda kingdom and was later adopted by missionaries for educational and religious dissemination during the colonial era.

Luganda is an agglutinative language, characterized by a complex system of morphemes that combine to form words. Unlike English's relatively fixed word order and analytic structure, Luganda employs a flexible syntax with extensive use of prefixes and suffixes to express grammatical relations and tenses. The language follows a Subject-Verb-Object order but is distinctive in its use of noun classes, which affect conjugations and adjectives, a feature not present in English.

The widespread use of Luganda is supported by its role in local media and education, where it is prevalent in radio, television, newspapers, and school curricula. It also functions as a lingua franca in urban areas and markets across different ethnic groups. The language's prominence is further bolstered by the Baganda, Uganda’s largest ethnic group, contributing to its extensive use and cultural significance in the region.

• Curated by: [Richard Kimera]
• Funded by [optional]: [No funding]
• Shared by [optional]: [Richard Kimera]
• Language(s) (NLP): [English, Luganda]

Dataset Sources [optional]

• Repository: [https://www.wordproject.org/bibles/]
• Paper [optional]: [More Information Needed]
• Demo [optional]: [More Information Needed]

Uses

This parallel corpus allows users to train translation models, develop language models for Luganda, and support cross-lingual learning for tasks like sentiment analysis and named entity recognition where Luganda data is scarce. It also facilitates linguistic research and comparisons between English and Luganda. Additionally, the dataset can be used for text summarization and part-of-speech tagging to enhance NLP capabilities in Luganda further.

Direct Use

Neural Machine Translation (NMT) experiments.

Out-of-Scope Use

Even though the datasets can build your NMT vocabulary, you have to be aware of the language context and the target usage of your model. The model trained should not be used for commercial purposes

Dataset Creation

The dataset was created by use of web-scraping techniques.

Source Data

The dataset was scrapped from https://www.wordproject.org/bibles/ for academic purposes.

Bias, Risks, and Limitations

This is the King James Version of the bible, and the language might not be used in everyday conversations. You might need a native speaker to improve the sentences to reflect the local context

Citation [optional]