Leveraging NLP and Machine Learning for Fine-Grained Emotion Detection from Textual Data

Abstract

New research into emotion detection from text is a quickly growing field of Natural Language Processing (NLP) that has many different applications, including mental health, customer service and social media analytics. This study looks at both traditional machine learning models and newer, more advanced deep learning models to help identify emotions, such as happiness, sadness, anger, fear and surprise. Baseline models (Logistic Regression, Naive Bayes, and Support Vector Machines (SVM)) are compared with three different deep learning architectures (recurrent neural networks (RNNs), long short-term memory (LSTM) networks, and the transformer BERT model). Labeled datasets (e.g., GoEmotions, CrowdFlower) are used to evaluate the accuracy, precision, recall and F1 score of the models. The findings indicate that transformer-based deep learning models (e.g., BERT) outperform baseline models and will likely play a crucial role in developing emotionally intelligent artificial intelligence systems.