Fine-Grained Emotion Recognition in Noisy Social Media Text Using Deep Bidirectional Sequence Models

Abstract

Social media has become an important medium for unbridled expressions of thoughts, opinions, and emotions in short textual formats. The sheer volume of textual content on these platforms makes it imperative to explore automatic emotion detection as an area of research in natural language processing and machine learning. This study aims to develop an efficient model for effective emotion detection based on a deep learning framework for multi-class emotion classification of short-form textual content on social media platforms. This study uses DAIR-AI Emotion Datasets containing tweets with emotion labels to train and test the model. This study uses a Bidirectional Long Short-Term Memory (BiLSTM) model to process contextual dependencies by considering both preceding and following word sequences to improve understanding of emotional contexts. The model aims to classify short-form textual content on social media platforms into six different emotions: sadness, joy, love, anger, fear, and surprise. Preprocessing techniques have been used to process the textual content before training and testing the model. The experimental results show that the proposed model based on the BiLSTM model achieves good classification results and effectively differentiates between emotions with subtle variations. This is evident from the analysis of the confusion matrices used to assess model performance. This study indicates that the proposed model is effective and practical for real-world applications. This study highlights the potential of deep learning models for effective emotion detection in noisy and informal communication on social media platforms.