Fake News Classification App with Sentiment Analysis

Overview

The Fake News Classification System is a Deep learning with machine learning Models, designed to identify and classify fake news articles accurately while having sentiment analysis to the news. Using cutting-edge Natural Language Processing (NLP) techniques and mahine and deep learning models, the application aims to combat misinformation and provide reliable tools for content verification.

---

Features

• Fake News Detection: Determines whether a given news article is authentic or fake.
• User-Friendly Interface: Simple and intuitive interface for submitting articles and viewing results.
• Real-Time Classification: Processes news articles quickly to provide classification results within seconds.
• Accuracy Tracking: Evaluates and tracks the model's performance using precision, recall, and F1 score.

---

Algorithms and Models Used

1. Logistic Regression:
   • Applied as an initial baseline model to compare results.
2. Natural Language Processing (NLP) Tools:
   • Tokenization, stemming, and lemmatization for preprocessing text data.
3. Convolutional Neural Network (CNN):
   • Used as the deep learning model for sequence classification tasks.
   • Captures spatial and hierarchical patterns in text data for better performance.

---

Functionality

1. Input News Text: Users provide the text of a news article they want to classify.
2. Preprocessing: The system tokenizes and cleans the input text for better model performance.
3. Prediction: The trained CNN-based model predicts whether the article is fake or authentic.
4. Output Results: The system displays the classification results along with a confidence score.
5. Performance Tracking: Ongoing evaluation to measure model improvements over time.

---

Technologies Used

• Python
• CNN
• Scikit-learn
• Pandas, Numpy
• Flask (for deploying a web application)

---

How to Use

1. Clone this repository.
2. Install the required dependencies using:

   pip install -r requirements.txt

3. Run the application:

   python app.py

4. Access the web application via http://localhost:5000.
5. Submit a news article for classification and view the results.

---

Future Improvements

• Add multilingual support for global usability.
• Incorporate real-time web scraping for news verification.
• Enhance model performance using ensemble learning methods.

---