Readers’ affect: predicting and understanding readers’ emotions with deep learning

Abstract Emotions are highly useful to model human behavior being at the core of what makes us human. Today, people abundantly express and share emotions through social media. Technological advancements in such platforms enable sharing opinions or expressing any specific towards others have shared, mainly form textual data. This entails an interesting arena for analysis; as whether there is a disconnect between writer’s intended emotion reader’s perception content. In this paper, we present experiments Readers’ Emotion Detection multi-target regression settings by exploring Bi-LSTM-based Attention model, where our major intention analyze interpretability effectiveness deep learning task. To conduct experiments, procure two extensive datasets REN-10k RENh-4k, apart from using popular benchmark dataset SemEval-2007. We perform two-phase experimental evaluation, first various coarse-grained fine-grained evaluations performance comparison with several baselines belonging different categories detection, viz., learning, lexicon based, classical machine learning. Secondly, evaluate readers’ detection assessing attention maps generated devising novel set qualitative quantitative metrics. The phase shows that Bi-LSTM + significantly outperforms all baselines. second analysis reveals may be correlated words well named entities.