Singular Value Decomposition Based Feature Extraction Technique for Physiological Signal Analysis

Abstract

Multiscale entropy (MSE) is one of the popular techniques to calculate and describe the complexity of the physiological signal. Many studies use this approach to detect changes in the physiological conditions in the human body. However, MSE results are easily affected by noise and trends, leading to incorrect estimation of MSE values. In this paper, singular value decomposition (SVD) is adopted to replace MSE to extract the features of physiological signals, and adopt the support vector machine (SVM) to classify the different physiological states. A test data set based on the PhysioNet website was used, and the classification results showed that using SVD to extract features of the physiological signal could attain a classification accuracy rate of 89.157%, which is higher than that using the MSE value (71.084%). The results show the proposed analysis procedure is effective and appropriate for distinguishing different physiological states. This promising result could be used as a reference for doctors in diagnosis of congestive heart failure (CHF) disease.

An eigenanalysis of data centering in machine learning

“…From this relation, it is easy to see that both matrices ∆ and ∆ ξ share the same eigenvectors, while any eigenvalue λ i of the former defines the eigenvalue ξ 2 λ i of the latter. By considering the normalization given in (16) with α i = 1 λi , we obtain from (29): X ⊤ w j = α j = X ξ ⊤ w ξ j . Therefore, projections onto axes defined by either PCA or ECA provide scale-invariant features, as show here within the MDS approach.…”

“…The intuitive motivation is the application of the spectral decomposition without data-centering in many pattern recognition and machine learning problems, thus providing a sort of a non-centered PCA by using the second-order non-central moment. This is the case for instance in signal analysis and classification [16] and in designing dictionaries for sparse representation [17]. A key motivation towards keeping data non-centered is the nonparametric density estimation with kernel functions [18], as revealed recently with exceptional performance in the (kernel) entropy component analysis (ECA) [19] and the information-theoretic learning framework [20].…”

A hybrid model for data visualization using linear algebra methods and machine learning algorithm

“…In recent years, there has been an exponential increase in the amount of digital information being generated across various fields, which has led to a significant surge in the size, complexity, diversity, and dimensions of data [1], which has given rise to a new type of data known as high dimensional data (HDD) [2], [3] HDD has been widely utilized across various industries, including healthcare, the Internet, education, commerce, and social networking [4], to name a few. The ever-increasing availability of new high-dimensional data can take on various formats, such as text [5], digital images [6], speech signals [7], and videos [8], among others. As such, developing new tools and techniques to manage and analyze such data effectively has become increasingly important to extract meaningful insights and drive innovation.…”