emoji-prediction

Emoji Prediction

Goal

Given a tweet, predict the most likely emoji that was used along with it.

Variants

Classification
Emoji Embeddings

MVPs

Classification

Train an embedding model (character or word) on the tweets.
Get tweet embeddings using averaging/any other method.
Classify tweets using KNN Naive Bayes.
Classify tweets using simple MLP.

Emoji Embeddings

Train an embedding model on the tweets along with the emojis.
Given a tweet without emoji, find its embedding (averaging of word embeddings).
Find closest emoji to it.

Taking it forward

Recurrent networks Train a simple single-layer RNN with softmax activation for classification. Use the embeddings obtained from the RNN model and see how they perform with MVPs.

Dataset

https://github.com/fvancesco/Semeval2018-Task2-Emoji-Detection/blob/master/dataset/Semeval2018-Task2-EmojiPrediction.zip?raw=true