Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{208560,
        author = {Pranav Chitalkar and Tushar Badgujar and Omkar Jadhav and Dr. Sunil Mahajan},
        title = {An AI-Powered Emotion-Aware Conversational Support System for Personalized Emotional Well-Being},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {no},
        pages = {482-494},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=208560},
        abstract = {Individuals frequently experience stress, anxiety, loneliness, and other emotional difficulties that often remain unaddressed because of limited access to counselling resources, social stigma, and hesitation to seek help in person. This paper extends an earlier rule-based sentiment analysis chatbot into a comprehensive AI-Powered Emotional Support and Mood Tracking Web Application that combines Natural Language Processing (NLP), Machine Learning (ML), and a personalized conversational interface. Rather than classifying a user's message only as positive, negative, or neutral, the proposed system analyzes free-form conversational text to infer sentiment polarity, a specific emotion category (such as sadness, loneliness, stress, anxiety, anger, or low motivation), and the intensity of that emotion, while also retaining conversational context across turns. The system produces empathetic, non-diagnostic responses, generates optional and non-intrusive self-care suggestions, and logs each interaction to build a longitudinal mood-tracking dashboard that helps individuals reflect on their own emotional patterns over time. A distinguishing feature of the proposed system is an optional, user-configured "Comfort Profile," through which a user may associate a photo of another person who is meaningful and comforting to them (male or female, such as a family member, friend, or mentor) along with that person's name, a favourite quote, and a preferred support style, so that the experience feels more personal and reassuring without the system ever implying that the depicted person is actually present or communicating. The architecture is designed as a full-stack application (React.js frontend, Node.js/Express backend, Python-based ML microservice, and MongoDB persistence layer) and proposes a staged model comparison between rule-based keyword matching, traditional machine learning (TF-IDF with Logistic Regression/SVM), and transformer-based models (BERT/DistilBERT), evaluated using accuracy, precision, recall, F1-score, and confusion-matrix analysis. A four-level escalation framework ensures that messages indicating severe distress or self-harm risk are handled through a clear, non-motivational safety pathway that directs individuals toward trusted contacts and professional or emergency support rather than continued casual conversation. The system is explicitly positioned as a supplementary, non-clinical, first-line emotional support and self-awareness tool, and not as a diagnostic instrument for depression or any other mental health condition.},
        keywords = {Sentiment Analysis; Emotion Detection; Natural Language Processing; Machine Learning; Mood Tracking; Context-Aware Chatbot; Mental Health; Personalized Conversational AI; MongoDB; Full-Stack Web Application},
        month = {September},
        }