This project focuses on the application of Multinomial Naive Bayes for the classification of fake news. The first part of the project involves classifying news articles into five classes (Barely-True, False, Half-True, Mostly True, Not-Known, True). The classification accuracy achieved in this part is relatively low, with only 23.77% correct predictions.
In the second part of the project, a different dataset with binary fake news classification (true/false) was used for classification. The results obtained in this case demonstrate a significant improvement, with 92.51% of the samples correctly classified.
The datasets used in this project can be found at the following links:
This project was implemented in July/August 2022 in the R language as part of the Advanced Statistics for Physics Analysis course offered by the University of Padova (academic year 2021-2022).
The project team consisted of:
-
- Tonazzo Valentina: https://github.com/ValentinaTonazzo
- Zoppellari Elena: https://github.com/zoppellarielena