Javascript must be enabled for the correct page display

Early Detection of Sepsis Induced Deterioration with First 48-hour ECG, Plethysmograph, and Respiratory Rate Biosignals Using Machine Learning Models

Dal Canton, Francesco (2018) Early Detection of Sepsis Induced Deterioration with First 48-hour ECG, Plethysmograph, and Respiratory Rate Biosignals Using Machine Learning Models. Bachelor's Thesis, Artificial Intelligence.

[img]
Preview
Text
AI_BA_2018_DalCantonF.pdf

Download (443kB) | Preview
[img] Text
toestemming.pdf
Restricted to Registered users only

Download (98kB)

Abstract

Sepsis is an excessive bodily reaction to an infection in the bloodstream, which causes 20% of patients to deteriorate within two days after hospital admission. Until now, no tool for early detection of sepsis induced deterioration has been found. This research uses ECG, respiratory rate, and blood oxygen saturation continuous bio-signals collected from 123 patients from the University Medical Center of Groningen during the first 48 hours after hospital admission. This data is examined under a range of feature extraction strategies and Machine Learning techniques as an exploratory framework to find the most promising methods for early detection of sepsis induced deterioration. The analysis includes the use of Gradient Boosting Machines, Random Forests, Linear Support Vector Machines, Multi-Layer Perceptrons, Naive Bayes Classifiers, and k-Nearest Neighbors classifiers. Promising results were obtained using Linear Support Vector Machines trained on features extracted from single heart beats using the wavelet transform and autoregressive modelling, where the classification occurred as a majority vote of the heart beats over multiple long ECG segments. None of the applied feature extraction strategies yielded better classification accuracies when paired with the tuned classifiers compared with the HRV measures extracted from the same dataset as part of the SepsiVit study, although features extracted with the wavelet transform and autoregressive modelling showed more promise.

Item Type: Thesis (Bachelor's Thesis)
Supervisor:
Supervisor nameSupervisor E mail
Wiering, M.A.M.A.Wiering@rug.nl
Quinten, V.M.v.m.quinten@umcg.nl
Degree programme: Artificial Intelligence
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 25 Jul 2018
Last Modified: 27 Jul 2018 12:44
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/18060

Actions (login required)

View Item View Item