Machine Learning project to detect fake news articles from text using FakeNewsNet dataset, and Google BERT algorithm.
We live in a post-truth world, where misinformation seems to increase all the time. We need a way to identify misinformation, apart from exhaustive, deep research on everything we read. Besides a few heuristics, can we go deeper with statistics and machine learning? This analysis shows that we can.
Main analysis is housed in file Capstone Project 2.ipynb
Custom modules for nlp and data processing are datascience.py and getnews.py
FakeNewsNet: https://github.com/KaiDMML/FakeNewsNet
Google BERT: https://github.com/google-research/bert