This project aims to classify emotions expressed in tweets using a deep learning model built with TensorFlow. Users can input text and receive predictions on the emotion conveyed.
- Preprocessing of tweet text data
- Deep learning model using TensorFlow and Keras
- Emotion classification into predefined categories
- Deployment using Streamlit for an interactive user interface
The dataset used for training the model contains tweets labeled with emotions such as:
- Anger
- Joy
- Sadness
- Fear
- Surprise
- Love
- Framework: TensorFlow
- Max sequence length: 50 (for text preprocessing)
- Embedding: Pretrained word embeddings or custom embeddings
- Architecture: LSTM/GRU-based deep learning model
- Loss Function: Categorical Crossentropy
- Optimizer: Adam
The model can be deployed using Streamlit to provide an interactive web-based interface for users to input tweets and receive emotion predictions in real-time.
This project was inspired by the Coursera course "Tweet Emotion Recognition with TensorFlow."
This project is licensed under the MIT License.