Emotion Recognition on Twitter using a Unison Model

Despite recent successes of deep learning in many fields of natural language processing, previous studies of emotion recognition on Twitter mainly focused on the use of lexicons and simple classifiers on bag-of-words models. In this paper whether we can improve their performance using deep learning and it is considered as the central question of the study. To this end, we exploit hash tags to create three large emotion-labeled data sets corresponding to different classifications of emotions. We then compare the performance of several word and character-based recurrent and convolutional neural networks with the performance on bag-of-words and latent semantic indexing models. By using this study we also investigate the transferability of the final hidden state representations between different classifications of emotions, and whether it is possible for predicting all of them using a shared representation to build a unison model. We show that recurrent neural networks, especially character-based ones, can improve over bag-of-words and latent semantic indexing models. The newly proposed training heuristic produces a unison model with performance comparable to that of the three single models even though the transfer capabilities of these models are poor.