Phenotypic Characterization of Chronic Kidney Patients Through Hierarchical Clustering

Annu Int Conf IEEE Eng Med Biol Soc. 2021 Nov:2021:2451-2454. doi: 10.1109/EMBC46164.2021.9630346.

Abstract

Chronic kidney disease is a major public health problem around the world and this disease early diagnosis is still a great challenge as it is asymptomatic in its early stages. Thus, in order to identify variables capable of assisting CKD diagnosis and monitoring, machine learning techniques and statistical analysis use has shown itself to be extremely promising. For this work, unsupervised machine learning, statistical analysis techniques and discriminant analysis were used.Clinical Relevance - Discriminating variables characterization assist to differentiate groups of patients in different stages of Chronic Kidney Disease and it has important outcomes in the development of future models to aid clinical decision-making, as they can generate models with a greater predictive capacity for Chronic Kidney Disease, predominantly aiding the early diagnosis capacity of this pathology.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Cluster Analysis
  • Humans
  • Kidney
  • Machine Learning*
  • Renal Insufficiency, Chronic* / diagnosis
  • Unsupervised Machine Learning