sdnjly/multisource-ecg-classification

Learning with Incomplete Labels of Multisource Datasets for ECG Classification

Learning with Incomplete Labels of Multisource Datasets for ECG Classification

This repository contains the source code and associated files for the research study titled "Learning with Incomplete Labels of Multisource Datasets for ECG Classification".

Table of Contents

• Introduction
• Getting Started
• Prerequisites
• Installation
• Data Preparation
• Usage

Introduction

This research study focuses on addressing the challenge of classifying electrocardiogram (ECG) data from multiple sources with incomplete labels. The study proposes a deep-learning-based framework to improve the accuracy of ECG classification in such scenarios. This repository contains the source code and resources related to the research.

Getting Started

These instructions will help you get a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

Before you begin, ensure you have met the following requirements:

• Python (>=3.6)
• TensorFlow (>=2.0)
• Other dependencies (listed in requirements.txt)

Installation

1. Clone the repository:

   git clone https://github.com/sdnjly/multisource-ecg-classification.git
   cd multisource-ecg-classification

2. Create a virtual environment (recommended):

   python -m venv venv
   source venv/bin/activate

3. Install required packages:

   pip install -r requirements.txt

Data Preparation

The datasets used in this study can be found from the Physionet/CinC challenge 2020/2021 websites. You can follow instructions on the websites to downloads the datasets.

• Physionet/CinC challenge 2020
• Physionet/CinC challenge 2021

Usage

1. model training

   python train_model.py training_data model

   training_data is a folder of training data files, model is a folder for saving your models, test_data is a folder of test data files

2. model testing

   python test_model.py model test_data test_outputs