Scholarlink

This chapter proposes a methodology to predict the Electrocardiogram (ECG) signals using two classes of software sensors. The first class is based on linear regression models such as Linear regression (LR), K-nearest neighbors (KNN), and Random Forest regression (RFR), this class used to forecast an ECG signal recorded in one auscultation site depending on another ECG signal registered at a different site. The second group contains time series forecasting models such as AR, ARMA and ARIMA, these models use the historical samples of an ECG signal to predict the future segment of the signal. The aim is to rebuild the missing data samples in case of any failure (power, sensor, technical, etc.). Using an actual database, the ability of the proposed software sensors to construct an ECG signal in case of data loss is discussed.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Software Sensor Design for Forecasting and Recovery of ECG Data

  • 章节

  • 发表日期：

    2026年

查看全文
  • Ghada Ben Othman,
  • Lilia Sidhom,
  • Inès Chihi,
  • Ernest Nlandu Kamavuako,
  • Mohamed Trabelsi

摘要

This chapter proposes a methodology to predict the Electrocardiogram (ECG) signals using two classes of software sensors. The first class is based on linear regression models such as Linear regression (LR), K-nearest neighbors (KNN), and Random Forest regression (RFR), this class used to forecast an ECG signal recorded in one auscultation site depending on another ECG signal registered at a different site. The second group contains time series forecasting models such as AR, ARMA and ARIMA, these models use the historical samples of an ECG signal to predict the future segment of the signal. The aim is to rebuild the missing data samples in case of any failure (power, sensor, technical, etc.). Using an actual database, the ability of the proposed software sensors to construct an ECG signal in case of data loss is discussed.