Figure - PMC

Skip to main content

An official website of the United States government

Here's how you know

Here's how you know

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

View full-text article in PMC

. 2020 Jan 17;27(3):437–443. doi: 10.1093/jamia/ocz211

Search in PMC
Search in PubMed
View in NLM Catalog
Add to search

© The Author(s) 2020. Published by Oxford University Press on behalf of the American Medical Informatics Association.

This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contactjournals.permissions@oup.com

PMC Copyright notice

Figure 3. — Feature selection results: (A) FS1 comprises high-profile hepatic and coagulation indicators (including bilirubin, Alanine transaminase [ALT], and the International Normalised Ratio [INR]); (B) FS2 features mainly comprise cardiovascular (troponin, lactate, d-dimer) and renal (creatinine, potassium, and urea) indicators; (C) FS3 features are relevant to all SOFA subcomponents, reflecting the heterogeneity of the minimal organ dysfunction; (D) FS4 features are of respiratory (CO2, CaO_2, and PO₂), and cardiovascular relevance (including mean arterial pressure [MAP]). (E) In contrast, FSS includes a portion FS1–FS4, but largely consists of general signs of deterioration (eg, systolic blood pressure, temperature, white blood cell count [WBC]), which have been found to be good indicators of mortality but nondiscriminative between septic and nonseptic patients.³¹ (F) shows the out-bag-error rates over 1000 iterations of bootstrapped feature selection.