kaw101010 / Telediagnosis.ai

Project for SpartaHack 8

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Telediagnosis.ai

Table of Contents

Description

An NLP-powered Disease Prediction System that allows user to express his symptoms through speech, extracts symptom, and returns a predicted ailment. Used GPT-3 for extracting symptoms from user's speech and used Random Forest Classifier to train the machine learning model.

Technologies Used

  • Used Python's Speech Recognition with Google to accept user's speech as input for natural language processing and utilised OpenAI API to access GPT-3 in order t0 extract symptoms from the text input
  • Used Python's scikit-learn library to train the model. The model was trained with a Random Forest algorithm.
  • Used HTML, CSS and JavaScript to develop the front-end and used Python Flask web framework for the backend.
  • Also used Flask to develop the API for the machine learning model

Setup

Clone and Fork this repository. Then navigate to the project directory. Then run pip install -r requirements.txt to install the required packages and dependencies. Then run python -m flask run to run the project with Flask.

Features

  • You can express your symptoms/problems by speech. Due to the efficient speech recognition library, your speech will be converted to text for symptom extraction.
  • Symptoms are being extracted by the most capable Davinci model of GPT-3.
  • The machine learning model is trained with Random Forest algorithm in the scikit-learn library.
  • The UI of the project is also simple and effective.

License

This project is licensed under the MIT License. More information about this license can be found here.

Contributors

Krish Wadhwani, Sanjay Marison, @ZilD117

About

Project for SpartaHack 8

License:MIT License


Languages

Language:Python 48.7%Language:CSS 22.7%Language:HTML 16.9%Language:JavaScript 11.7%