Powered by Machine Learning algorithms, simple web-app to help find early stage diabetes prediction.
Check out the app @ https://diabetes-classifier-qech.onrender.com
Using the power of three different machine learning classifiers, this website can help predict early stage diabetes.
This project was designed and built during Software Engineering Semiar course, and was my first project in the realm of Machine Learning.
You can find the full seminar report (in Hebrew) at this link.
The models used in this projects are: Perceptron, Multilayer Perceptron and Decision Tree.
The database used to train the models was taken from Kaggle and can be found here.
The project is written in Python with the website powered by Flask.
The models object is taken from the scikit-learn library.
The web app is hosted on heruko.
*Please note that when first loading the website it may take up to 30 seconds to load due to heruko free-tier hosting service.
To run the web app on your machine, clone this repo https://github.com/birkagal/diabetes-classifier.git
Open terminal window at the project directory an run
python app.py
The website would be available at https://localhost:5000
- Fix decision tree problem caused on heruko. (Currently decision tree is disabled.)
There are more than 450,000 diabetics living in Israel who are treated with medication.
according to estimates by the National Diabetes Council there are another 200,000 people with undiagnosed and untreated diabetes.
Diagnosis of diabetes is important to treat the disease and prevent the complications associated with diabetes.
To use the website, answer ALL of the following questions. The answers will be used in a Machine Learning model I created
and the result would appear on screen.
PLEASE NOTE!: This site is used for learning porpusus only
and if you think you have diabetes,
PLEASE SEEK PROFFESIONAL DOCTOR.
Gal Birka - birkagal@gmail.com
Distributed under MIT License. See LICENSE for more information.