AN INTUITIVE APPRAOCH ON LSTM-HYBRID MODEL FOR MULTI-DISEASE PREDICTION AND CHATBOT RECOMMENDATION SYSTEM

Abstract

In the realm of healthcare, accurate disease classification is essential for effective diagnostics and treatment planning. Traditional methods often struggle with the complexity and variability of medical data. This paper introduces a novel approach by combining user query responses with a hybrid Long Short-Term Memory (LSTM) model that integrates Convolutional Neural Networks (CNNs) and a chatbot-based recommendation system. User-generated queries are processed using natural language processing (NLP) techniques and then analyzed by the hybrid LSTM-CNN model, which captures both sequential and contextual dependencies to enhance classification accuracy. The CNN enhances feature extraction, while the LSTM manages long-term dependencies, improving the precision and recall of the disease classification. The chatbot not only interacts with users to provide tailored recommendations and information but also aids in gathering contextual data that feeds into the classification model. Evaluation on a diverse dataset shows that this combined approach significantly outperforms traditional methods, offering a scalable, context-aware solution for multi-disease classification and demonstrating substantial potential for advancing healthcare outcomes through sophisticated machine learning techniques and interactive systems.