Fake News Prediction with Logistic Regression

Introduction

This repository contains the code for a Fake News prediction system using Logistic Regression. The code is based on a Jupyter notebook originally generated by Colab.

Functionality

This code performs the following tasks:

• Import Libraries: Imports necessary libraries like pandas, numpy, nltk etc. for data manipulation, text processing and machine learning.
• Data Preprocessing:
  • Loads the training data (train.csv) into a pandas dataframe.
  • Handles missing values by replacing them with empty strings.
  • Combines author name and title into a single "content" column.
  • Separates the data (content) and the target label (fake/real).
  • Applies stemming to reduce words to their root form and removes stopwords (common words like "the", "and").
  • Converts textual data into numerical features using TF-IDF vectorizer.
• Train-Test Split: Splits the data into training and testing sets for model evaluation.
• Model Training: Trains a Logistic Regression model on the training data.
• Evaluation:
  • Evaluates the model's accuracy on both training and testing data.
• Prediction:
  • Makes a prediction on a new unseen piece of text data (example from the testing set).
  • Classifies the news as Real or Fake based on the prediction.

Running the Code

This code is intended to be run in a Jupyter Notebook environment. You can follow these steps:

1. Download the code and data files.
2. Open the Fake_News_Prediction.ipynb file in a Jupyter Notebook environment.
3. Run the code cells sequentially.

Dependencies

• Python 3.x
• pandas
• numpy
• nltk
• scikit-learn

l