Skip to content

Generic fake news detector across news domains, using feature engineering. Fully implemented POS tagging, sentiment analysis, emotion detection, topic modelling and named entity recognition.

Notifications You must be signed in to change notification settings

jermynyeo/fake-news-detection

Repository files navigation

Fake News Detector

Medium Article

User Guide

  1. Compiled.ipynb is to show the whole flow of our analysis (without output). The process is as follow :
  2. EDA and Preprocessing is the exploration to understand our dataset prior to analysis and creation of dataframes of the policitcs dataset
  3. TopicModelling is used to generate the topic distribution for both politics and entertainment domain
  4. Emotion_Sentiment_Pos is used to genereate the emotion distribution, sentiment distribution and pos grouping for politics domain
  5. Named Entity Recognition is used to generate the entity type count for Location, Persons and Organisations
  6. CrossDomain_Emotion_Sentiment_POS is used to genereate the emotion distribution, sentiment distribution, pos grouping and entity type count for entertainment domain that will be used for testing for cross domomain model
  7. Classification shows the model training , testing and validation process. It also shows the type of models used and the respective data used for both specific and cross domain classifiers
  8. process_text is to process the dataset from the entertainment dataset.

About

Generic fake news detector across news domains, using feature engineering. Fully implemented POS tagging, sentiment analysis, emotion detection, topic modelling and named entity recognition.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published