Study on ECG Classification Based on PCA and SRC

  • Dechun Zheng, Guangchun Gao

Abstract

This paper described an arrhythmia classification algorithms based on the technologies of Principal Component Analysis (PCA) and sparse representation classifier (SRC) for the purpose of heartbeat recognition. Using the MIT-BIH arrhythmia classification database, heartbeat sample data adopted in this paper are segmented from ECG records. To increase the classification capability, the heartbeat waveform is clustered by using k-means clustering algorithms. After data is preprocessed, according to these clustering data, the characteristic matrix is established based on Principal Component Analysis. Finally, combined with PCA, the model of the proposed classification method is obtained by using SRC. 600 ECG samples are investigated, of which 500 were training samples and 100 were test samples. The result of the experiment shows that the recognition rate of the same patient is 97%, and the recognition rate of different patients is 88%.

Published
2020-10-31
How to Cite
Dechun Zheng, Guangchun Gao. (2020). Study on ECG Classification Based on PCA and SRC. Design Engineering, 336 - 344. https://doi.org/10.17762/de.vi.774
Section
Articles