Paper Title
Comparing Deep Learning Techniques for Emoticon Prediction in Hindi Sentences
Abstract
Emoticons are small graphics that are commonly used in texting on social media platforms. They provide a
mixture of graphical and textual content at the same time and offer a unique manner of connection. The analysis of
emoticons allows a better understanding of the user’s feelings and thus is an important part of understanding natural
language. In the research presented below, we analyzed the correlation between emoticons and words belonging to the Hindi
language and predicted the appropriate emoticons for a collection of text or Twitter messages. We used sequence-tosequence
learning with word embeddings and encoder-decoder-based machine translation to translate the text to English and
then utilised various machine learning and deep learning based approaches to predict the emoticons. We present a
comparison of the various techniques we used during our experiments.
Keywords – Emoticon Prediction, Neural Machine Translation, BiLSTM, GRU