Description
To tackle the problem of fake news, a chrome extension is developed that can be used to predict fake news articles. The API used by this extension is developed by training several news articles using Natural Language Processing techniques. The news articles are ultimately categorized into biased, fake or satire. The model is deployed in a docker container hosted on Azure VM.
How to install
- Extract the zip folder
- Go to chrome://extensions/
- Click 'Load unpacked'
- Select the extracted folder
How to use
- Go to any news source.
- Click on the extension icon
- The score represents the prediction score between 0 and 1. Higher the score, more authentic is the news source
- The category represents whether the domain is to be trusted or not
- The decision represents whether the URL is predicted to be impartial or biased
Examples
- Fake example: http://wearethellod.com/breaking-rnc-office-that-held-clinton-evidence-burned-to-the-ground/
- Satire example: https://www.huzlers.com/man-hospitalized-after-his-apple-airpods-exploded-in-his-ear/
- Not fake: https://sputniknews.com/science/201903021072895578-mars-underground-water-system-evidence/
Log in or sign up for Devpost to join the conversation.