Patient Health Status Recognition Based on Electrocardiogram and Neural Network
Autor: | Križanec, Matija |
---|---|
Přispěvatelé: | Šimunić, Dina |
Jazyk: | chorvatština |
Rok vydání: | 2018 |
Předmět: |
RBBB
srce neural network ventricular bigeminy TEHNIČKE ZNANOSTI. Računarstvo neuronska mreža prijevremena ventrikularna kontrakcija scaled conjugate gradient automatic classification electrocardiogram ventrikularna tahikardija elektrokardiogram TECHNICAL SCIENCES. Computing automatska klasifikacija skalirani konjugirani gradijent heard ventricular tachycardia |
Popis: | U ovom radu prikazan je sustav za automatsku klasifikaciju EKG signala u klase: normalan sinusni srčani ritam, prijevremena ventrikularna kontrakcija, blokada desne provodne grane i ventrikularna tahikardija. Digitaliziran EKG zapis preuzet je iz MIT-BIH baze te je nad uzorcima od 28 pacijenata izvršeno učenje, testiranje i validacija neuronske mreže, koja je pokazala zadovoljavajuće rezultate. Iako klasifikacija koju klasifikator nudi nije opsežna, omogućila je uvid u mogućnosti koje neuronske mreže pružaju nad svakodnevnim problemima ljudskog života. In this paper, a system for automatic classification of ECG signals is modeled and developed. It can classify given ECG signals into four classes: normal sinus rhythm, ventricular bigeminy, right bundle branch block and ventricular tachycardia. Digitalized ECG record is taken from MIT-BIH database and over the samples of 28 patients neural network was trained, tested and validated. The statistical analysis has shown satisfactory results. Although, the classification of the classifier is not extensive, it has shown the possibilities of neural networks and the changes it can make in everyday life. |
Databáze: | OpenAIRE |
Externí odkaz: |