Tensor-based approximation of multichannel ECG sections

Abstract

Tensors are multi-dimensional arrays which can be used to represent multi-dimensional data. Here we show how to approximate and compress multiple QRS-aligned multichannel ECG sections, each comprising P-wave, QRS-complex, and T-wave, by higher order singular value decomposition (HOSVD). We also introduce a method for the selection of components obtained by HOSVD for data approximation and a measure to quantify the quality of this tensor-based approximation of multi-dimensional data. The compression performance of this lossy data compression approach is assessed by the compression ratio.