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{205820,
        author = {Sania Navgire and Pranjali Nalawade and Faheem Manyar and Dr. Gajanan P Arsalwad},
        title = {HelixCare AI: NLP-Based HealthCare Advisor and Emergency Support System},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {1},
        pages = {7957-7966},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=205820},
        abstract = {This paper presents HelixCare AI, a web-based conversational system that delivers preliminary, informational guidance on common infectious diseases through natural language interaction. The system accepts free-form symptom descriptions, identifies the most probable intent from a fixed catalogue of disease- and symptom-related categories, and returns a corresponding advisory response. The core classifier is a feedforward neural network operating over a Bag-of-Words representation of user input, where each query is tokenized, stemmed using the Porter algorithm, and converted into a fixed-length binary vector prior to classification. The network consists of three fully connected layers with rectified linear unit (ReLU) activations, trained using the Adam optimizer and cross-entropy loss. At inference time, a softmax confidence score is computed for the predicted intent, and a fixed threshold of 0.75 is applied to decide whether to return a disease-specific response or a fallback message indicating that the query was not understood. The system is implemented using Django, PyTorch, and NLTK, with the dataset stored in a structured JSON format covering fifty-three infectious-disease and health-advisory categories. This paper documents the architecture, dataset characteristics, train-ing procedure, and empirical performance of the system, and discusses the implications and limitations of threshold-based, intent-driven healthcare chatbots. The system is intended strictly for informational and preliminary advisory purposes and does not replace professional medical diagnosis.},
        keywords = {Bag-of-Words, chatbot, feedforward neural net-work, healthcare advisory system, intent classification, natural language processing.},
        month = {June},
        }