/DiabetO

This was mostly done during a 24 hour hacakathon which took place in VIT-AP on February 15th of 2024.

Primary LanguageJupyter NotebookMIT LicenseMIT

hax-feb-15

Repository to collaborate for the hackathon taking place on Feb 15, 16 in VIT-AP.

Problem statement (AI): . Deep Learning in Healthcare: Create a model that predicts patient health outcomes based on electronic health records

Diabetes Detection System by Team Dard Inc.

Table of Contents

  1. Team Members
  2. Project Inspiration - Daredevil
  3. Project Description
  4. Key Features
  5. Technical Challenges
  6. Project Impact
  7. Unique Aspect - X-Factor
  8. How to use
  9. License

Team Members

  • Herbert George (22BCE7969)
  • Adithya Bijoy (22BCE9270)
  • Bhumit Chaudhry (22BCE8331)
  • Syed Qasim Mustafa (22BCE7974)
  • Neha Sharma (22BCE8513)

Project Inspiration - Daredevil

Our project draws inspiration from Daredevil, a superhero known for his commitment to helping others. Much like Daredevil's enhanced senses, our AI model focuses on detecting diabetes using various health attributes, such as BMI levels, age, and blood pressure.

Project Description

Our team has developed a machine learning algorithm that analyzes electronic health records to predict the presence of diabetes in patients. The model considers factors like the number of pregnancies, glucose levels, blood pressure, skin thickness, insulin levels, BMI levels, diabetes pedigree function, and age. By deriving comprehensive data from patients' health records, the model creates plots to enhance predictive accuracy.

Key Features and Technical Challenges

  • Key Features

    • Utilizes machine learning to predict diabetes based on multiple health attributes.
    • Achieves an accuracy level of approximately 80% after rigorous training and testing.
    • Addresses challenges in data cleaning and visualization for effective model performance.
    • User-friendly interface for quick and easy predictions using inputted data.

  • Technical Challenges

    • Cleaning and preprocessing the dataset to ensure quality input for the model.
    • Visualization of data to enhance understanding and accuracy in predictions.

Project Impact

Our diabetes detection system contributes to the improvement of the healthcare sector by providing a faster and cost-effective preliminary test for diabetes. Hospitals can leverage this model to expedite the detection process, leading to early interventions and reduced testing costs.

Unique Aspect - X-Factor

Our model stands out by utilizing a well-established machine learning algorithm that considers eight crucial factors, resulting in increased accuracy and reliability. The user interface further distinguishes our project, offering a quick and easy prediction process for user-inputted data.

How to Use

  1. Clone the repository.
  2. Install the required dependencies.
  3. Run the application and input the necessary health attributes.
  4. Receive quick and accurate predictions regarding the likelihood of diabetes.
  5. Or... diabeto.onrender.com/static diabeto.onrender.com/flet

License

This project is licensed under the MIT License - see the LICENSE file for details.