SEARCH RESULT

The rapid advancement of artificial intelligence in healthcare has opened new avenues for early disease prediction and clinical decision support. This paper presents a Multi-Disease Prediction System (MDPS) that integrates machine learning and deep learning models for predicting various diseases using sequential data for better understanding of each patient’s health conditions. Specifically, this research extends previous studies and addresses the challenges and limitations. Many open-source models and systems are available but all have very generic datasets and based on machine learning Deep learning can make models more complex but also gives more useful insights but it needs different types of datasets and processing. Based on the data collected for each disease, we selected the most suitable model some are deep learning and some are machine learning depending on what works best. Some models are left as machine learning because they already perform well and cannot be improved much with current data. the results of this investigation are relatively surprising since previous studies have mainly focused on machine learning classification for most of the part. LSTM / RNN and DNN had a significant impact on the temporal (continuous) data.to understand the feature relationship, we have tried to implement GNN based model which gives insights which can be mainly used by the hospitals for the analysing of the key feature relationship and impact of them on patient’s Health. This hybrid, modelspecific approach offers valuable support for healthcare professionals. The system also includes a simple health chatbot for basic health-related conversations and suggestions that is currently powered by ollama models but can configure any NLP models based on availability. A user-friendly Streamlit web app is used as the frontend to make the system easily accessible for both users and healthcare professionals. Streamlit integrates all the models and provides a compact website with a stacked prediction system in one place. This can be a life-changing solution in rural areas where diagnostic materials and tools are not easily available and for normal users who want to check their health conditions by themselves.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Rishiram B Aravind A Vinay Vunnava Kanipriya M