streamlit-ecg

A simple Streamlit app that classifies heartbeats in order to detect Arrhythmia.

Description

This app is based on the "MIT-BIH Arrhythmia Database" dataset avaiable on Physionet and is able to :

Download the full dataset
Extract the records and the beats
Train a SVC model
Show the prediction in a web app

Create your .env file

All your settings are stored in that file. Please create a new .env file in the root folder and edit your variables.

Here is an example of the .env file :

#Path
DB_PATH="data/mit-bih-arrhythmia-database-1.0.0"
CACHE="streamlit-ecg/cache"

#Signal
MIN_BPM=35
MAX_BPM=185
FILTER=0
RESAMPLING=100
WORKERS=3

#Model
RANDOM_STATE=15

#Window
DELTA=10

Run the projet

The streamlit app is launch with the following command

streamlit run streamlit-ecg/stream.py

Open your web browser and go to <LOCALHOST>:<8501> to see the streamlit app.

Model

Score : 0.965

Credits