Fake News Detection of Twitter Data using Passive-agressive Classifier. I also have made WebApp deployed using Flask & Heroku.
Datasets : "politifact_real.csv" & "politifact_fake.csv"
After analysis Data I got to know:
Percentage of Real & Fake Data are kind of same.(50%)
Wordcloud of Frequent words of Fake Data:
Wordcloud of Frequent words of Real Data:
I have also done sentiment analysis.
we can clearly see fake datasets have negative polarity.
Screenshot of Heroku app deployment -->"fake-news-1o7"
DEMO: