GDS_TwoTime

Data

To reproduce the results from the report, you need to download the dataset 995,000_rows.csv.

If the data/ directory does not exist in the root of the repository, create it by running:

mkdir -p data

Place the 995,000_rows.csv file inside the data/ directory.

note to contributers

.gitignore file is already configured to ignore everything inside the data/, you can verify that the file is ignored by running:

git check-ignore -v data/995,000_rows.csv

for data documentation see

https://github.com/several27/FakeNewsCorpus

Required Libraries

The script requires the following Python libraries:

1. nltk - Natural Language Toolkit for text processing.
2. pandas - Data manipulation and analysis.
3. numpy - Numerical computing.
4. sklearn (scikit-learn) - Machine learning utilities
5. matplotlib - Plots.
6. cleantext - Corpus Cleaning
7. pandarallel - paralisation
8. gensim
9. joblib
10. clean-text

Pre-trained models

• If you're within the institution of UCPH, you can access a repostoire of our models, pre-trained and ready for testing. This includes the Doc2Vec models, which can take a long time to train otherwise. The link is below: https://alumni-my.sharepoint.com/:f:/g/personal/mrs530_alumni_ku_dk/EoXl2FTCYhtAha9IFT6XmG4BEJvQKtrX2t61GiMmLc8Byw?e=zVcP8w

Installation Instructions

To install the required libraries, ensure you have Python installed on your system. You can install the dependencies using pip by running the following command in your terminal or command prompt:

pip install nltk pandas numpy scikit-learn matplotlib cleantext pandarallel