The Cyberbullying Warning System is an AI-driven solution designed to detect and address cyberbullying in online communications. By leveraging Natural Language Processing (NLP) and Machine Learning (ML), this system analyzes text inputs for toxicity and harassment indicators, providing proactive warnings to users.
Cyberbullying has become a significant concern in the digital age, leading to emotional distress and psychological harm. This project aims to mitigate such effects by identifying harmful content and alerting users before it spreads.
- Automated Text Analysis: Uses NLP techniques such as tokenization, stemming, and sentiment analysis to process textual data.
- Machine Learning Models: Implements Logistic Regression, Random Forest, and Support Vector Machines (SVM) to classify text based on toxicity levels.
- Active Learning Approach: Enhances dataset quality dynamically through iterative labeling and annotation.
- User-Friendly Interface: Integrates with MS Excel and Tableau for intuitive content analysis and visualization.
- Improved Accuracy: Achieved a 30% accuracy improvement over baseline models through hyperparameter tuning and model selection.
- Programming Language: Python
- Libraries & Frameworks:
- NLP: NLTK, SpaCy
- Machine Learning: Scikit-Learn, TensorFlow
- Data Visualization: Tableau, Matplotlib, Seaborn
- Data Handling: Pandas, NumPy
- Successfully analyzed diverse online comments for toxicity indicators.
- Developed a scalable system that can be extended for use in social media monitoring and online forums.
- Reinforced the importance of AI-driven solutions for promoting online safety.
- Integration with social media platforms for real-time monitoring.
- Expansion of the dataset to improve model generalization.
- Implementation of deep learning approaches for enhanced detection accuracy.