Loading...

Single-replicate longitudinal data analysis in the presence of multiple instrumental measurement errors

Moazeni, M ; Sharif University of Technology | 2020

262 Viewed
  1. Type of Document: Article
  2. DOI: 10.1016/j.cie.2020.106301
  3. Publisher: Elsevier Ltd , 2020
  4. Abstract:
  5. In this paper, a novel method as a combination of the expectation–maximization (EM) algorithm and Variogram is proposed to decompose the longitudinal measurement errors in the absence of replications and the presence of multiple instrumental measurement errors. In the proposed method, multiple measurements are considered where the units are observed by several distinct instruments (gauges). The approach decouples the observed variance of the measurement model into the process and measurement system variances. In addition, it decomposes the variance of multiple instruments into the process and instrument variances. In the end, the proposed model is validated and tested based on simulated longitudinal data as well as a real case study related to the Framingham Heart study, measuring systolic blood pressure by multiple instruments. In addition, the robustness of the proposed method to the missing values, a common problem in longitudinal data, is demonstrated. © 2020 Elsevier Ltd
  6. Keywords:
  7. EM-algorithm ; Reproducibility ; Variogram ; Blood pressure ; EM algorithms ; Longitudinal data ; Missing values ; Multiple instruments ; Reproducibilities ; Variograms ; Measurement errors
  8. Source: Computers and Industrial Engineering ; Volume 141 , 2020
  9. URL: https://www.sciencedirect.com/science/article/abs/pii/S0360835220300358