You are on page 1of 15

Fake news classification using NLP

Team Members

S.NO Roll Number Name

1 20X41A4217 G.Nikhitha

2 20X41A4235 M. Abdul Naveed

3 20X41A4237 N S Praneetha Gandikota

4 20X41A4248 S.Sri Divijendra Kumar


Abstract

• Deceptive content pretending to be real news is a global issue that impacts


people's opinions, decisions, and voting. Fake news often starts on social
media and spreads to mainstream media. A study on identifying fake news
shows a new detector using linguistic features like quoted attribution to
estimate the likelihood of a story being fake. The detector is 63.333%
accurate at identifying fake news articles with quotes. This process, called
influence mining, can help detect fake news and propaganda. The paper
discusses the research process, analysis, linguistics, and detector
performance, as well as how this system may evolve in the future.
Keywords

• Misleading Words
• Bias and Subjectivity
• Source Reliability
• Inflammatory Language
• Social Media Indicators
• Contextual Features
Introduction

• Fake news classification using Natural Language


Processing (NLP) is a critical task in combating
misinformation. By employing advanced linguistic
analysis, NLP algorithms discern patterns, emotional
language, and reliability indicators to differentiate
between authentic and deceptive news, aiding in
the preservation of accurate information in an era of
digital misinformation.
Problem Statement

The challenge is to develop robust Natural Language


Processing (NLP) solutions that can effectively discern
between authentic and deceptive information
Requirements

• Software Requirements:
• Python IDE (3.7 version), Anaconda (3.7), Jupyter
• Hardware Requirements:
• Minimum: Intel i3 processor, 4GB RAM, 250GB
hard disk
• Operating Systems: Windows, Linux
Model and Prediction
Model

• Dataset - Preprocessing , NLP Tools


• Feature Extraction - Word2Vec,Bag of Words
• Supervised Learning Algorithms - Binary
Classification
• Model Architecture
• Training
• Model Check
Binary Classification Algorithms

• There are quite a few different algorithms used in


binary classification. The two that are designed
with only binary classification in mind (meaning
they do not support more than two class labels)
are Logistic Regression and Support Vector
Machines. A few other algorithms are: Nearest
Neighbours, Decision Trees, and Naive Bayes.
Prediction

• Input Processing
• Embedding Generation
• Prediction
• Threshold Setting
• Real-Time Processing
Conclusion
• This structured approachs leads to a
robust model that can differentiate
between real and fake news articles,
addressing the important issue of
misinformation in the digital world.
Thank you

You might also like