Overview
Key Features
- News API Integration: Authenticates tweets against Times of India news API for fact-checking
- Keyword Matching: Filters tweets based on contextual similarity with reference news articles
- Sentiment Analysis: Determines if tweets are negative or assertive using NLP techniques
- XOR Logic: Applies logical operations to detect contradictions between tweets and verified news
- Configurable Similarity: Adjustable SimilarityDegreeValue for keyword matching precision
- Misinformation Detection: Identifies tweets that contradict verified news sources
How It Works
Step 1: Contextual Filtering
- Filters tweets that are out of context by matching keywords with reference news
- Uses configurable SimilarityDegreeValue to determine relevance
- Ensures only relevant tweets are analyzed
Step 2: Sentiment Analysis
- Applies NLP functions to detect negative operations in text
- Classifies tweets as either Negative or Assertive
- Uses pattern recognition to identify sentiment patterns
Step 3: XOR Operation
- Performs XOR operation between tweet sentiment and news sentiment
- Binary classification: 0 for negative, 1 for positive
- Detects contradictions: If tweet (0) XOR news (1) = 1, the tweet contradicts the news
Step 4: Misinformation Identification
- Flags tweets where XOR result is 1 (contradiction detected)
- Identifies false information that contradicts verified news sources
- Provides actionable results for fact-checking
Technologies Used
- Python: Core programming language
- NLTK (Natural Language Toolkit): Natural language processing and text analysis
- Pattern: Advanced pattern matching and sentiment analysis
- News API: Times of India API for authenticated news sources
- Twitter API: For fetching and analyzing tweets
Technical Highlights
- Implemented sophisticated NLP algorithms for text analysis
- Created configurable similarity matching system
- Developed XOR-based contradiction detection logic
- Integrated multiple APIs for comprehensive fact-checking
- Built scalable architecture for processing large volumes of tweets
Challenges and Solutions
Challenge 1: Contextual Relevance
Solution: Implemented keyword matching with configurable similarity thresholds, allowing fine-tuned control over what constitutes relevant content.
Challenge 2: Sentiment Classification
Solution: Used NLTK and Pattern libraries for advanced sentiment analysis, detecting negative operations and classifying text accurately.
Challenge 3: Contradiction Detection
Solution: Implemented XOR logic to mathematically detect contradictions between tweet sentiment and news sentiment.
Real-World Applications
- Social Media Monitoring: Automated detection of misinformation on Twitter
- Fact-Checking Services: Supporting human fact-checkers with automated analysis
- News Verification: Cross-referencing social media claims with verified news sources
- Research: Academic research on misinformation patterns and detection methods
Outcome
- Proficiency in Python and NLP libraries (NLTK, Pattern)
- Understanding of sentiment analysis and text classification
- Ability to design algorithms for real-world problems
- Integration of multiple APIs and data sources
- Problem-solving skills in combating misinformation