Fake News

Description

The purpose of this project is to analyze a Kaggle data set and to create a predictive model to label fake news.

Data

The Kaggle data provided by Clément Bisaillon.

EDA

A pandas-profiling report is available.

Code

All code is available in the following folder as Jupyter notebooks.

• Code with bias mitigation
  • Notebook Viewer
• Accuracy Training using 6 months
  • Notebook Viewer
• Accuracy Training using 1 year
  • Notebook Viewer
• Original code
  • Notebook Viewer

Documentation

Executive Report

Presentation

Executive Presentation

Instructions

To run this notebook locally, install Jupyter, download the data set, change the file location to load the code and data, and install all the library dependencies.

Try Anaconda.

Model Folders

• Pickle Format
• PyCaret Format

Tools

• Python
• Jupyter Notebook
• Pandas-profiling
• Scikit-learn
• Pycaret
• Tableau
• MS PowerPoint
• Neattext

Credits

News image by Nijwam Swargiary at Unsplash