In today’s world where short messages and tweets are at the core of communication, emojis have become major forms of expression of ideas and emotions. They penetrate language barriers and allow people to express a whole lot in a very concise manner. With the increasing use of emojis in our daily life, sometimes we lose context of text and aren’t really sure about which emoji to use based on the text. Our project aims to suggest emojis based on the given text by analyzing the sentiment of the given text and predicting relevant emojis for it.
- Python >= 3
- NLTK >= 3.2.3
- Keras >= 2.0.7
- Word Embeddings
- Download here
- Place them in the src folder
- Naive Bayes Classifier (
nb) - Decision Tree Classifier (
dtc) - LSTM (
lstm) - Bi-Directional LSTM (
blstm)
Clone the repo.
Navigate to the src directory.
Run deploy_model with your choice of model as an arg.
Eg.
python3 deploy_model nbFor further knowledge of the implementation, strategies and accuracies received, please refer to this file.
The project is available under the MIT License. Check the license file for more information.