Overview

This repository hosts the source code for three distinct machine learning models, each fine-tuned for specific Natural Language Processing (NLP) tasks:

1. Text Emotion Recognition: Classifies the emotional content in a given text.
2. Sentiment Analysis: Analyzes text to determine the sentiment (positive, negative, neutral).
3. Cyberbullying Detection and Classification: Identifies and categorizes instances of online harassment or bullying.

Model Performance Metrics

After extensive training and fine-tuning, each model has demonstrated strong performance as indicated by the following metrics:

Text Emotion Recognition

Model Architecture: Fine-tuned RoBERTa
Performance Metrics:

• Accuracy: 92.8%
• Precision: 93%
• Recall: 93%
• F1 Score: 93%

Sentiment Analysis

Model Architecture: Fine-tuned RoBERTa
Performance Metrics:

• Accuracy: 95.76%
• Precision: 96%
• Recall: 96%
• F1 Score: 96%

Cyberbullying Detection and Classification

Model Architecture: Fine-tuned XLNet
Performance Metrics:

• Accuracy: 99.56%
• Precision: 99.56%
• Recall: 99.56%
• F1 Score: 99.56%

License

The project is licensed under the MIT License.