Fake News Prediction Model

This project implements a Fake News Prediction Model using Logistic Regression. The goal is to classify news articles as either real or fake. This is a basic binary classification task.

---

Overview

With the increasing spread of misinformation, this model aims to detect fake news using machine learning techniques. It leverages text preprocessing methods such as stemming and vectorization to prepare the data for model training. The model uses a logistic regression algorithm, which is suitable for binary classification.

---

Dataset Description

The dataset used for this project is from Kaggle: Fake News Dataset. It contains both real and fake news articles, which have been labeled accordingly.

Features:

• id: unique id for a news article
• title: he title of a news article
• author: author of the news article
• text: the text of the article; could be incomplete
• label: a label that marks the article as potentially unreliable (0 = Real, 1 = Fake)

---

Model Performance

The model performed well on both the training set and testing set, achieving the following accuracy:

• Training Accuracy: 98.64%
• Testing Accuracy: 97.91%

Due to time constraint the model was only trained on an attrbute combination of 'author' and 'title' rather than 'text'

Usage

Here’s how to use the model to predict whether a given news article is real or fake:

# Example usage with test data
X_new = X_test[1]  # Select a news article from the test set

# Predict whether the news is real or fake
prediction = model.predict(X_new)

# Print the result
if prediction == 1:
    print('THIS NEWS IS FAKE!')
else:
    print('THIS NEWS IS REAL!')