Andishe_ye Amari

Andishe_ye Amari

Analysis of censored survival data using adequate dimensionality reduction methods: The Tehran Lipid and Sugar Study

Document Type : Original Article

Authors
1 Master's degree in Statistics, Shahid Beheshti University, Tehran, Iran
2 Faculty of Statistics, Shahid Beheshti University, Tehran, Iran
3 Zeshk Senior Researcher, Metabolic Disease Prevention Research Center, Endocrine and Metabolic Sciences Research Institute, Shahid Beheshti University of Medical Sciences, Tehran, Iran
4 Researcher, Metabolic Disease Prevention Research Center, Endocrine and Metabolic Sciences Research Institute, Shahid Beheshti University of Medical Sciences, Tehran, Iran 5
5 Epidemiologist, Metabolic Disease Prevention Research Center, Endocrine and Metabolic Sciences Research Institute, Shahid Beheshti University of Medical Sciences, Tehran, Iran
Abstract
Cardiovascular diseases are the most common cause of death worldwide. On the other hand, to determine an appropriate survival model to predict the risk of developing heart diseases and identify important risk factors in the occurrence of these diseases, it is necessary to determine the functional form that relates survival time and risk factors. In this study, a sufficient dimension reduction method using a general model that includes common survival models as special cases is proposed to predict the risk of developing heart diseases.
Sufficient dimension reduction methods based on inverse regression combined with the Cox proportional hazards model have a good predictive performance for future survival of individuals.
Keywords

Volume 23, Issue 2
February 2019
Pages 17-25

  • Receive Date 10 May 2025
  • First Publish Date 10 May 2025
  • Publish Date 20 February 2019