Human emotions are a critical aspect of intelligent human–computer interaction. Conventional chatbots generally focus on intent recognition and fail to capture the emotional context behind user inputs, which limits their usefulness in sensitive and interactive applications. In this work, we present an AI-based emotional detection chatbot that identifies user emotions from textual communication using natural language processing and machine learning techniques. The system classifies emotions such as joy, sadness, anger, fear, love, and neutrality, and generates empathetic and context-aware responses in real time. A transformer-based emotion classification model is combined with text auto-correction and synonym-based emotion mapping to improve robustness against informal language and spelling variations. Additionally, the proposed framework supports real-time interaction through WebSocket communication and is designed to be extensible for live video-based facial emotion analysis. Experimental observations show that the chatbot provides emotionally relevant responses and improves user engagement compared to traditional sentiment-based systems. The proposed system demonstrates the potential of emotionally intelligent conversational agents in applications such as mental health support, education, and customer interaction platforms.
Gowthami et al. (Tue,) studied this question.