Fake-News-Detection

This project explores the use of supervised learning models to detect fake news with a 7796×4 dataset (source: Data Flair). The dataset can be viewed here: https://drive.google.com/file/d/1er9NJTLUA3qnRuyhfzuN0XUsoIC4a-_q/view

News.csv format:

Column 1 | Column 2 | Column 3 | Column 4 ID | Title | Text | Label ("Real"/"Fake")

Outputs

Accuracy of 5-fold Cross Validated Logisitic Regression

Accuracy of 5-fold Cross Validated Passive Aggressive Classifer

Train Test Split on Logit and PAC

Improvements

Some potential next steps include:

• improve data quality for learning
• test additional models for performance
• introduce additional attributes/factor columns