Digitala Vetenskapliga Arkivet

Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Emergency Department Triage Prediction of Emergency Severity Index using Machine Learning Models
KTH, School of Electrical Engineering and Computer Science (EECS).
KTH, School of Electrical Engineering and Computer Science (EECS).
2019 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesisAlternative title
Akutmottagningens förutsägelse av Emergency Severity Index med hjälp av Maskininlärningsmodeller (Swedish)
Abstract [en]

Study Objective: The emergency department (ED) in the United States strongly rely on subjective assessment of patients. This study seeks to evaluate an electronic triage system based on machine learning models that can predict the patients emergency severity index (ESI).

Methods: A dataset containing 560 486 patients triage data was investigated.Three different machine learning models was tested and evaluated. A crossvalidation table and a confusion matrix was conducted from each of the models. The precision rate, recall rate and f1-score were calculated and reported.

Result: The Gradient Boosting model returned an accuracy rate of 68%. The random forest model returned an accuracy rate of 66%. The Gaussian Naive Bayesmodel returned an accuracy rate of 25%.

Conclusion: The model that best predicted the ESI-level is the GradientBoosting model. Further testing is needed with better computational power since we could not train our model with the whole dataset.

Abstract [sv]

Syfte: Akutmottagningen i USA förlitar sig kraftigt på en subjektiv värdering av patienter. Denna studie söker efter att evaluera ett elektronisk triage systembaserad på maskininlärningsmodeller som kan förutse patienters ESI.

Metod: Ett data set som innehåller 560 486 patienters triage data har undersökts. Tre olika maskininlärningsmodeller har testats och evaluerats. En cross validation tabell och en confusion matrix har skapats för varje modell. Precision, recall och f1 värde har kalkylerats och rapporterats.

Resultat: Gradient Boosting modellen har returnerat ett accuracy värde av 68%. Random Forest modellen har returnerat ett accuracy värde av 66%. Gaussian Naive Bayes modellen har returnerat ett accuracy värde av 25%.

Slutsats: Modellen som har bäst förutsett ESI nivåerna är Gradient Boostingmodellen. Flera tester behövs med starkare beräkningskraft då vi inte kunde träna vår modell med hela datasetet.

Place, publisher, year, edition, pages
2019. , p. 34
Series
TRITA-EECS-EX ; 2019:341
Keywords [en]
Machine Learning, Triage, ESI, Classification
Keywords [sv]
Maskininlärning, ESI, Klassificering
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:kth:diva-259402OAI: oai:DiVA.org:kth-259402DiVA, id: diva2:1351508
Educational program
Master of Science in Engineering - Computer Science and Technology
Supervisors
Examiners
Available from: 2019-09-16 Created: 2019-09-16 Last updated: 2022-06-26

Open Access in DiVA

fulltext(827 kB)723 downloads
File information
File name FULLTEXT01.pdfFile size 827 kBChecksum SHA-512
961e4508d3567d1f41682dfccfa51df682cb26db48935ebc0ded6d3691b4f360f3962e7207228667ce1448abc3cac89c6ec6cb9e6acfc437a64bc86448dcef47
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Computer and Information Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 723 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 1384 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf