by Gergő BOGNÁR and Sándor FRIDLI
Department of Numerical Analysis, Faculty of Informatics, ELTE Eötvös Loránd University
Budapest, Hungary, 2018
|
|
Description
MATLAB code of our classification framework ECG Heartbeat Classification by Means of Variable Rational Projection (to be published).
Requirements, dependencies
- MATLAB with Parallel Computing Toolbox, Signal Processing Toolbox and Wavelet Toolbox
How to install and use
- Download and extract the framework to a desired folder.
Step 2 and 3 can be skipped if you downloaded the full version including the toolboxes and the database.
- Download the MIT-BIH Arrhythmia Database and set the
G_MITDB
environment variable to its path in g_init.m
(default: data/mitdb
).
- Download the WFDB, RAIT, and LIBSVM toolboxes and add their path to MATLAB in
g_init.m
(default: libs
).
- Set the
G_OUTPUT
environment variable to the desired output folder in g_init.m
(default: results
).
All intermediate and final results will be saved to this folder, including the converted and preprocessed database, the extracted features, the classification results, and the final fusion results.
- The framework can be invoked with the
g_ecg_classification
command. If you only want to try one part then call it directly or set the G_ENTRY
environment variable in g_init.m
(default: 0).
We note, that the third, classification part can take multiple hours to execute due to the high amount of data.
Download
- Source code as a ZIP archive (30,7 KB).
- Full code with the toolboxes and the database included (72,6 MB).