Synthetic News Generation for Fake News Classification

This dataset contains synthetic news articles and tweets generated using fact-based manipulation techniques as described in "Synthetic News Generation for Fake News Classification" (Sittar et al.). The dataset includes 338 synthetic news articles and 481 synthetic tweets created using the Llama3.1:8B large language model, with original content focused on COVID-19 vaccination topics.

The synthetic content was generated through a structured methodology that:

Extracts critical facts from authentic articles

Systematically modifies these facts

Regenerates coherent content with altered factual elements

Each synthetic item maintains the stylistic properties of the original while containing deliberately modified factual information, making this dataset valuable for research on fake news detection, content verification systems, and the evaluation of misinformation classification models.