This Android application predicts ICU readmissions using discharge summaries. The application supports two methods for obtaining discharge summaries:

1. FHIR Server Integration: Fetch discharge summaries directly from a hospital's FHIR server.
2. OCR Integration: Capture a discharge summary document using the device camera, extract the text using OCR, and predict ICU readmission.

The backend server processes discharge summaries and predicts ICU readmission probabilities using a pre-trained machine learning model.

• Main Screen:
  • Navigate to either FHIR Server or OCR-based prediction.
• FHIR Integration:
  • Enter the patient ID to fetch the discharge summary from the hospital's FHIR server.
  • Predict ICU readmissions based on the retrieved discharge summary.
• OCR Integration:
  • Capture an image of the discharge summary.
  • Perform text recognition using Google ML Kit's OCR.
  • Predict ICU readmissions based on the extracted text.

• FHIR Server Integration:
  • Fetch discharge summaries using the FHIR API.
• Prediction Model:
  • A pre-trained machine learning model (Linear SVC) is used to predict ICU readmissions.
  • The model uses TF-IDF vectorization to process the text data.
• Text Cleaning:
  • Text is cleaned by removing punctuation, extra spaces, and normalizing case before prediction.

• Frontend:
  • Android Studio installed on your system.
  • Device or emulator with internet access for testing the app.
• Backend:
  • Python 3.8+
  • Required Python libraries (Flask, fhirpy, pickle, scikit-learn, etc.).