Free dataset for fake news detection

Here are some free datasets for fake news detection:

  1. LIAR Dataset: This dataset is a collection of labeled statements from the TV show "The Daily Show with Jon Stewart" that were fact-checked by PolitiFact. The dataset contains 12,836 statements, with 5,551 labeled as "true" and 7,285 labeled as "false". Source: Cornell University
  2. Fakenewsnet Dataset: This dataset contains 12,000 news articles, with 6,000 labeled as "fake" and 6,000 labeled as "real". The dataset is collected from various sources, including online news websites and social media platforms. Source: Fakenewsnet
  3. FakeNews Dataset: This dataset contains 1,000 news articles, with 500 labeled as "fake" and 500 labeled as "real". The dataset is collected from online news websites and social media platforms. Source: Kaggle
  4. SciFact Dataset: This dataset contains 1,000 scientific articles, with 500 labeled as "true" and 500 labeled as "false". The dataset is collected from online scientific journals and websites. Source: Stanford University
  5. FakeNews-500 Dataset: This dataset contains 500 news articles, with 250 labeled as "fake" and 250 labeled as "real". The dataset is collected from online news websites and social media platforms. Source: GitHub
  6. NewsVerification Dataset: This dataset contains 1,000 news articles, with 500 labeled as "verified" and 500 labeled as "unverified". The dataset is collected from online news websites and social media platforms. Source: University of California, Berkeley
  7. FakeNewsDetection Dataset: This dataset contains 1,000 news articles, with 500 labeled as "fake" and 500 labeled as "real". The dataset is collected from online news websites and social media platforms. Source: GitHub

Please note that these datasets may have varying levels of quality, and some may require additional preprocessing or cleaning before use. Additionally, some datasets may have restrictions on their use or require attribution. Be sure to review the terms and conditions of each dataset before using it.