Summary
- Developed a deep learning-based solution to classify emotions in English Twitter messages using Natural Language Processing (NLP) techniques.
- Conducted data pre-processing, tokenization, and feature extraction to improve model accuracy and performance.
- Implemented multiple learning models collaborated with a team, including LSTM and BERT, ensured a clear understanding and alignment across the team.
- Delivered detailed reports and analysis on model’s metrics, effectively presented project concept and outcomes to the sponsors.
This project aimed to develop a deep learning-based solution for classifying emotions in English Twitter messages using advanced Natural Language Processing (NLP) techniques. The team implemented and compared multiple models, including Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Convolutional Neural Networks (CNN).
The Hugging Face dair-ai Emotion dataset, consisting of English Twitter messages labeled with six basic emotions, served as the foundation for model training and evaluation. A combination of data pre-processing, tokenization, and feature extraction techniques was employed to improve the performance of these models, ensuring robust and reliable emotion classification.
