FakeNews is a Python command-line application and library for building an ensembled set of classifiers to detect fake news articles. This is a work in progress, based on basic understandings of classification problems in relation to NLP. A first approach will attempt to create a feature space using n-grams from extracted text, select a feature subset via TF-IDF, and build a classifier using random forest.
- Install apt packages: sudo apt-get install python3-dev openjdk-7-jre libxml2-dev libxslt1-dev libssl-dev libffi-dev
- Create virtualenv: virtualenv --python=python3.4 .venv
- Activate virtualenv: . .venv/bin/activate
- Install pip requirements: pip install -r requirements.txt
- Install python-boilerpipe: https://github.com/misja/python-boilerpipe
The following is a list of research papers on fake news detection:
The following is a list of text extraction library evaluations: