AI Augmented Mental Health Journal with Sentiment Analysis and Mood Visualization Using NLP and LLM

Abstract

Stress, anxiety, and emotional instability are becoming more common today, but getting professional psychological help is still hard because of cost, stigma, and lack of availability. Recent progress in Natural Language Processing (NLP) and Large Language Models (LLMs) makes it possible to build smart systems that help people think about their emotions and improve their health. This paper introduces MoodMate AI, a web application for journaling and emotional analysis that uses AI to combine sentiment analysis, fine-grained emotion classification, semantic embeddings, and LLM-driven empathetic response generation. The system employs a hybrid pipeline combining rule-based sentiment analysis (VADER), transformer-based emotion classification, Sentence-BERT (SBERT) embeddings, and a generative LLM for personalized feedback. A scalable web architecture enables real-time interaction, interpretability, and modular extensibility. Experimental evaluation demonstrates effective sentiment detection, promising emotion recognition performance, and low-latency inference, validating the feasibility of deploying emotionally intelligent AI systems for real-world well-being applications.