qianyxxx / ml-fake_news Goto Github PK

Simple NLP and ML project to improve applied ML skills with a real-world problem

Will be predicting whether a headline and article text represents real or fake news from words that appear. For this purpose, implemented Naive Bayes algorithm from scratch without using any external Machine Learning library and used Bag-of-words model with different parameters. Also used Tf-idf to extract features for comparison.

Fake News is a dataset provided to determine when an article is fake or real. It includes the following features:

• id: unique id for a news article

• title: the title of a news article

• author: author of the news article

• text: the text of the article, could be incomplete

• label: a label that marks the article as potentially unreliable

    1: unreliable
    0: reliable
  

• Öner İnce