We are solving the problem of fake COVID19 news levitation during such epidemic time. Rightnow due to inavailability of past data of COVID19 (as it is a new aspect for all), the previous fake news detection algorithms outperform. Here we propose a more accurate ML model to the detect, segregate COVID19 news from other news and to pin them are real or fake.
Our platform feature include ML model 1. Scrapping news articles from various sources 2. Differentiating Covid news from others 3. Finding whether news is fake or not 4. Extracting Regions
Android and Web Extension An Android and Browser portable extension, where in user can simply pin any news, posts that he thinks maybe fake. Such posts are feed to our ML model and output is generated. Also all such posts can be diverted to Designated Responsible Operator, who can pin such news into fake or real news.
Android and Web App This App will be user friendly app which will show precise stats and news and general purpose important information to the end user. Also a non-covid more which can be used to get positive useful news, posts in this pandemic time
Setting up intermediate bot, for example : A WhatsApp bot can be set up, if user thinks that a post is fake then he can just forward this news to our bot and get the correct result.
Contraints : Proper labelled dataset for fake news detection and malicious URL detection.
Known issues : 1. Platforms like whatsapp are end to end encrypted, enriching such dataset is difficult 2. For extracting regions from the text, it does not classifies the regions not in our dataset(due to inavailability of proper dataset for regions in India).
A detailed solution is included in the attached PDF : https://drive.google.com/file/d/1mbFQq3aJqAUHzsYiqeY6mYgMaaUIbvz5/view?usp=sharing
Try It out
android, bootstrap, css, flask, html, java, javasrcipt, kotlin, materialui, mongodb, newsapi, nltk, python, scikit, scraping, selenium, spacy, tensorflow, web