Stanford: Using Neural Networks to Predict Emoji Usage from Twitter Data

Stanford: Using Neural Networks to Predict Emoji Usage from Twitter Data. “As emojis have emerged as an increasingly important and standardized part of modern textual inputs, analyzing their usage patterns are becoming of great interest to any modern text system. We frame this investigation as a text classification problem, mapping input text to their most likely accompanying emoji, utilizing 1.5 million scraped Twitter tweets for our training set. Both our LSTM-RNN model and our CNN model outperform our baseline, although the CNN model surprisingly achieves much better accuracy and F1 results. We conclude the paper by proposing future works in this area.”