NLP Project Report: Textual Emotion-Cause Pair Extraction in Conversations

Abstract

In the dynamic field of computational linguistics, understanding emotional triggers in conversations is crucial for developing empathetic AI systems. This paper introduces a novel model designed to identify the causes of emotions within conversational contexts. Leveraging interdisciplinary approaches from psychology and advanced natural language processing (NLP), our model integrates attention mechanisms, transformers, and deep learning-based convolutional neural networks (CNNs). It uniquely incorporates the Myers-Briggs Type Indicator (MBTI) to analyze speakers’ personality traits and behavioral patterns. This comprehensive approach allows for precise predictions of emotional triggers, answering the fundamental question: ”Why do we feel a certain way during conversations?” Our findings have significant implications for understanding and provides a approach to incorporate emotional intelligence in modern chatbots and AI-driven communication systems.