LiberationKG: A Knowledge Graph of Bangladesh Liberation War from Newspaper Articles
Mohammad Fahim Foisal, Rudra Pratap Deb Nath, Nasrin Sultana · 2024
The Bangladesh Liberation War of 1971 remains one of the most significant and emotional events in the country’s history. In today’s digital world, historical narratives are at risk of being distorted by disinformation and the spread of false news. Newspaper articles from unbiased sources provide a reliable overview of the situation. However, while unstructured texts are easily understandable for humans, they are not readily interpretable by machines. To address this, we present Libera-tionKG, a knowledge graph of the Bangladesh Liberation War, generated from reputable sources such as Time Magazine and The New York Times. LiberationKG is created by producing a total of 12,291 Resource Description Framework (RDF) triples from the mentioned sources using a Large Language Model and constructs from RDFS, OWL, and Schema.org vocabularies. Its development involves advanced natural language processing techniques, RDF triple validation and optimization, and linking to external resources. LiberationKG provides an organized representation of knowledge and an efficient system for information extraction. With 80% accuracy, LiberationKG is linked to 521 external resources and adheres to FAIR principles, ensuring interoperability and data accessibility, making it a valuable digital resource for preserving this pivotal historical event.