Skip to content

Latest commit

 

History

History
45 lines (23 loc) · 1.86 KB

README.md

File metadata and controls

45 lines (23 loc) · 1.86 KB

Multiple Disorder Prediction App

Welcome to the Multiple Disorder Prediction app! This application predicts the likelihood of obesity and diabetes in a person based on various inputs. It utilizes machine learning models, pipelines, and column transformers to efficiently handle data and provide predictions.

Overview

This project incorporates machine learning models trained to predict the probability of obesity and diabetes in individuals. The models are developed in a Jupyter Notebook environment using pipelines, column transformers, and exported as pickle files for easy deployment.

Technology Stack

  • Model Development: Jupyter Notebook
  • Machine Learning Algorithms: Logistic Regression, Descision Tree Classifier, implemented using pipelines and column transformers
  • Frontend: Streamlit
  • Deployment: Streamlit Cloud Services

Getting Started

Clone the Repository

To run the application locally, clone this repository using the following command:

git clone https://github.com/ayushtiwari134/multiple_disorder_predictor

Running the App

After cloning the repository, navigate to the project directory and execute the following command to run the app:

streamlit run app.py

This command will start the Streamlit web application locally, enabling access to the multiple disorder prediction interface.

Deployment

The application is deployed using Streamlit Cloud Services, offering a live environment to predict the likelihood of obesity and diabetes in individuals.

Features

  • Input Parameters: Users can input various health-related factors, such as BMI, blood sugar levels, age, etc.
  • Prediction: The application predicts whether a person is obesity and/or diabetic based on the provided inputs.
  • Efficient Data Processing: Utilizes pipelines and column transformers for efficient data handling and model predictions.