Identification of myocardial infarction using consumer smartwatch ECG measurement
Kivi, Samuel (2023-05-05)
Identification of myocardial infarction using consumer smartwatch ECG measurement
(05.05.2023)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
avoin
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe2023051042729
https://urn.fi/URN:NBN:fi-fe2023051042729
Tiivistelmä
The goal of this thesis is to detect and classify acute myocardial infarctions from smartwatch ECG data. As the smartwatches have been increasing in numbers, and many of new smartwatch models have capability to detect ECG data. This study aims to answer to the question whether or not the ECG data from smartwatches can be used to detect acute myocardial infarctions.
To answer to this question, and existing database has been used in tandem with smartwatch ECG data gathered from two different smartwatches. Five different machine learning models have been used to detect and classify ECG data. The best performing machine learning model was Extra Trees, which achieved accuracy of 90.84% with using Leave-One-Out Cross-Validation.
These results show that ECG data from smartwatches could be used to detect infarctions. Measuring ECG with smartwatch is much easier than using clinical ECG measurement devices, meaning that ECG measuring could reach much wider audience that it has prior to this been able to reach.
Further research could include gathering larger database from smartwatch ECG, and the data ownership of smartwatch, and other medical and biological data that companies collect.
To answer to this question, and existing database has been used in tandem with smartwatch ECG data gathered from two different smartwatches. Five different machine learning models have been used to detect and classify ECG data. The best performing machine learning model was Extra Trees, which achieved accuracy of 90.84% with using Leave-One-Out Cross-Validation.
These results show that ECG data from smartwatches could be used to detect infarctions. Measuring ECG with smartwatch is much easier than using clinical ECG measurement devices, meaning that ECG measuring could reach much wider audience that it has prior to this been able to reach.
Further research could include gathering larger database from smartwatch ECG, and the data ownership of smartwatch, and other medical and biological data that companies collect.