Regresar

Processing of the electrical signal of heart, using machine learning

Abstract:

Blood pressure PA is one of the most important factor in analyzing the cardiac system, along with the heart's electrical signal or electrocardiographic signal SECG biological variables, these variables allow know to some extent the performance of the human cardiac system. Today there are few studies that show the relationship between the two variables and that demonstrate the relationship between PA and SECG. It's paper shows that there a relationship between PA and SECG. This means that you can find the values of PA using the SECG. For this study, the sampling was carried SECG, 22 patients, 18 healthy between 17 and 26 years and 4 with altered PA 50 and 78, for this the Powerlab equipment was used, where electrodes were used signal to capture the heart through the DI, once obtained the sample signals, the composition of this was studied, especially R and T waves in order to have individual zones and find the values of systolic and diastolic blood pressure, using a system of artificial intelligence such as neural networks and support vector machines. We used Wavelet Transform for segment the SECG in its R and T wave, and shows its equivalent in systole and diastole portions. These signals and found patterns were used as training the neural network and support vector machines, was used supervised learning. In this paper we showed that was possible to find the PA of a patient using neural networks and support vector machines with a total success rate of 98.488 to 95,708 for systole and diastole respectively. Finally, a mobile application is used to display the values of PA in real time.