KGolemo / fake-news-detection

Recognition of so called ‘fake news’ on specified topic scope - BSc Thesis (5.0) - Dataset of 2240 messages related to coronavirus, data vectorization based on word uni- and bigrams and the TF-IDF method, classification carried out with the use of a naive Bayes classifier, a linear support vector machine and a random forest algorithm.

Recognition of so called 'fake news' on specified topic scope

BSc Thesis

Usage

1. Data acquisition from the FakeHunter portal - jupyter-notebooks/web_scraping_fakehunter.ipynb
2. Data acquisition from the Termedia portal - jupyter-notebooks/web_scraping_termedia.ipynb
3. Data processing - jupyter-notebooks/data_preprocessing.ipynb
4. Feature extraction and application of classifiers - jupyter-notebooks/modeling.ipynb

Documentation

The documentation is available here.

Web app

Web app is available here. Its repository can be found here.