Diabetes Monitoring Dashboard

This is a healthcare web-application that uses Machine Learning algorithms to predict whether a person is diabetic or not, while also providing val…

15 lines🖥 Desktop only

View source folder on GitHub →contributed by

@Akarsh3053

Diabetes Monitoring Dashboard

This is a healthcare web-application that uses Machine Learning algorithms to predict whether a person is diabetic or not, while also providing valuable life style improvement suggestions through chatGPT api.

How is it unique and different from other diabetes prediction models ?

It uses only the information from the user which he/she can fill in real-time.
This application is integrated with an IOT device responsible for measuring real-time blood glucose levels.
App evaluates the data and returns not only the prediction but also a good suggestion on how to keep up a good fitness for every individual.

Chat-GPT is already integrated in the dashboard to help user with further insights regarding results.😊

Model Description

We used SVM(Support Vector Machine) machine-learning algorithm on the dataset that was uploaded to the repository (diabetesv2.0). This dataset contains a minimum number of parameters that are required to do predictions for diabetes.

The used machine learning model in this project is SVM at 77.27% accuracy.

Here are some performance metrics for our model:

SVM Accuracy: 0.7727272727272727
SVM Precision: 0.7272727272727273
SVM Recall: 0.5818181818181818
SVM F1 Score: 0.6464646464646464

Requirements

Install the till-needed packages using the command :

pip install -r requirements.txt

Also, don't forget to add your own OenAI API-key in chat.py

How to run the application

After installing all the dependencies, open a terminal window in project directory and run following command :

streamlit run webApp.py

The application will deploy a webapp on localhost which then can be accesed through web browsers (Chrome recommended!) by any client on that network.

Pyodide-runnable

No — it runs a Streamlit web server, makes network requests, uses scikit-learn, and calls the OpenAI API.