Unsupervised classification of 12-lead ECG signals using wavelet tensor decomposition and two-dimensional Gaussian spectral clustering
作者：
Highlights：
• Developed a new feature extraction approach of 12-lead ECG signals in tensor space based on DWPT and MPCA.
• Realized the tensorization of ECG samples and the dimensionality reduction of ECG wavelet tensors in a tensor space.
• Devised a new hybrid distance measure for constructing the proximity matrix of spectral clustering.
• Developed a novel two-dimensional Gaussian spectral clustering for 12-lead ECG signals.
• The practical lab dataset and two datasets from PhysioBank are used to verify the efficiency of the proposed method.
摘要
•Developed a new feature extraction approach of 12-lead ECG signals in tensor space based on DWPT and MPCA.•Realized the tensorization of ECG samples and the dimensionality reduction of ECG wavelet tensors in a tensor space.•Devised a new hybrid distance measure for constructing the proximity matrix of spectral clustering.•Developed a novel two-dimensional Gaussian spectral clustering for 12-lead ECG signals.•The practical lab dataset and two datasets from PhysioBank are used to verify the efficiency of the proposed method.
论文关键词：12-lead ECG signal,Spectral clustering,Dimensionality reduction,Wavelet packet transform,Tensor
论文评审过程：Received 14 February 2018, Revised 28 August 2018, Accepted 2 September 2018, Available online 15 October 2018, Version of Record 21 November 2018.
论文官网地址：https://doi.org/10.1016/j.knosys.2018.09.001
