<?xml version="1.0" encoding="UTF-8"?><?xml-stylesheet type="text/xsl" href="static/style.xsl"?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2026-09-30T04:06:40Z</responseDate><request verb="GetRecord" identifier="oai:skemman.is:1946/39409" metadataPrefix="dim">https://skemman.is/oai/request</request><GetRecord><record><header><identifier>oai:skemman.is:1946/39409</identifier><datestamp>2021-06-22T13:52:57Z</datestamp><setSpec>com_1946_6870</setSpec><setSpec>com_1946_6001</setSpec><setSpec>col_1946_34329</setSpec></header><metadata><dim:dim xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:doc="http://www.lyncode.com/xoai" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.dspace.org/xmlns/dspace/dim http://www.dspace.org/schema/dim.xsd">
<dim:field mdschema="dc" element="contributor" lang="is">Háskólinn í Reykjavík</dim:field>
<dim:field mdschema="dc" element="contributor" qualifier="author">Benedikt Hólm Þórðarson 1993-</dim:field>
<dim:field mdschema="dc" element="description" qualifier="advisor">Michal Borsky 1986-</dim:field>
<dim:field mdschema="dc" element="description" qualifier="advisor">Jacqueline Clare Mallett 1963-</dim:field>
<dim:field mdschema="dc" element="description" qualifier="abstract" lang="is">Vaxandi samstaða vísindamanna á sviði svefnrannsókna er að kæfisvefn sé röskun sem hrjáir stóran hluta fólks. Kæfisvefn veldur fjölda neikvæðra heilsufarslegra áhrifa, þar á meðal einkenna eins og mikla syfju yfir daginn. Til að greina kæfisvefn eru sjúklingar settir í svefnrannsókn. Svefnmerkjafræðingur fer þá handvirkt yfir upptökuna og setur inn merkingar eftir settum stöðlum. Þær merkingar sem eru gerðar eru að lokum notaðar til að ákvarða alvarleika kæfisvefnsins sem hrjáir sjúklingurinn. Þessi ritgerð kynnir aðferð til sjálfvirkrar flokkunar á kæfisvefni í einstökum andardráttum. Aðferðin nýtir aðeins 3 merki sem ekki eru ífarandi í mælingu og er almennt safnað í svefnrannsókn. Þessu er náð með því að merkja fyrst einstaka andardrætti og þar á eftir þjálfa ýmis sjálfvirk vélanáms módel til að skipta andardráttum í flokka&#xd;
eftir kæfisvefni. Kynnt er nýtt reiknirit til að greina andardrætti og árangur þess metinn. Reikniritið nær 0.94 F1 stigi þegar greindur er andardráttur sem gefur til kynna framúrskarandi greiningartíðni. Með því að nota andardráttinn sem greindur er með reikniritinu eru ýmis tölugildi sem lýsa andardrætti skilgreind og metin, með það að&#xd;
markmiði að ákvarða hæfi þeirra til að aðgreina andardrætti sem gerast í kæfisvefni. Að lokum eru tölugildin notuð til að þjálfa sjálfvirk módel í að greina andardrætti, annars vegar sem kæfisvefn eða ekki. Mest nákvæmni náðist með HMM módeli sem gat greint andardrætti með 95% nákvæmni.</dim:field>
<dim:field mdschema="dc" element="description" qualifier="abstract" lang="en">An emerging consensus in the field of sleep research is that obstructive sleep apnea is a disorder afflicting a large subset of the population. The disorder has a multitude of negative health effects, including symptoms such as excessive daytime sleepiness. To correctly diagnose obstructive sleep apnea, patients undergo a sleep study. A trained&#xd;
sleep technician must then manually score the recording according to scoring rules. The resulting scoring is finally used to determine an apnea-hypopnea-index, which is an indicate of the disorder’s severity.&#xd;
This thesis introduces a method for automatic classification of obstructive apnea in individual breaths which utilises only 3 non-invasive signals that are commonly collected in a sleep study. This is achieved by first segmenting the signals on a breath-to-breath&#xd;
basis, and then training various automatic classifiers to categorise breaths into apneic or non-apneic categories. A novel breath segmentation algorithm is introduced, and its performance is evaluated. The algorithm achieves a very high classification accuracy with an F1 score of 0.94 score when detecting breaths, which indicates an excellent detection rate. Using the breaths detected by the segmentation algorithm, various intra-breath features are then defined. The features are evaluated with respect to their suitability to separate breaths that occur during obstructive apnea from nonobstructive ones. Finally, the features are used to train automatic classifiers to classify breaths into either obstructive or non-obstructive categories. The highest accuracy was achieved by a hidden Markov model classifier, which was able to detect obstructive apnea breaths with 95% accuracy.</dim:field>
<dim:field mdschema="dc" element="description" qualifier="sponsorship" lang="is">Nox Medical, Íslensk erfðagreining (deCODE Genetics), and Rannís (Icelandic Research Fund Rannís grant #175256-0611)</dim:field>
<dim:field mdschema="dc" element="date" qualifier="accessioned">2021-06-22T13:52:55Z</dim:field>
<dim:field mdschema="dc" element="date" qualifier="available">2021-06-22T13:52:55Z</dim:field>
<dim:field mdschema="dc" element="date" qualifier="issued">2021-06-22T13:52:56Z</dim:field>
<dim:field mdschema="dc" element="date" qualifier="submitted">2021-06-11T14:31:55Z</dim:field>
<dim:field mdschema="dc" element="date" qualifier="published">2021-06</dim:field>
<dim:field mdschema="dc" element="identifier" qualifier="uri">http://hdl.handle.net/1946/39409</dim:field>
<dim:field mdschema="dc" element="language" qualifier="iso">en</dim:field>
<dim:field mdschema="dc" element="subject" lang="is">Tölvunarfræði</dim:field>
<dim:field mdschema="dc" element="subject" lang="is">Meistaraprófsritgerðir</dim:field>
<dim:field mdschema="dc" element="subject" lang="is">Reiknirit</dim:field>
<dim:field mdschema="dc" element="subject" lang="is">Svefnrannsóknir</dim:field>
<dim:field mdschema="dc" element="subject" lang="is">Kæfisvefn</dim:field>
<dim:field mdschema="dc" element="subject" lang="en">Computer science</dim:field>
<dim:field mdschema="dc" element="subject" lang="en">Computer algorithms</dim:field>
<dim:field mdschema="dc" element="subject" lang="en">Sleep apnea syndromes</dim:field>
<dim:field mdschema="dc" element="subject" lang="en">Research</dim:field>
<dim:field mdschema="dc" element="title" lang="en">Analysis and detection of obstructive apnea in individual breath cycles</dim:field>
<dim:field mdschema="dc" element="title" lang="is">Greining kæfisvefns í einstökum andardráttum</dim:field>
<dim:field mdschema="dc" element="type">Thesis</dim:field>
<dim:field mdschema="dc" element="type" qualifier="degree">Master's</dim:field>
</dim:dim></metadata></record></GetRecord></OAI-PMH>