Research On Denoising Algorithm Of Dynamic ECG Signal Based On Supervised Learning Depth Factor Analysis

Nowadays,cardiovascular disease has been the hotspots in the medical researches with the high morbidity,high mortality,and high recurrence rate.In the context of telemedicine,Holter monitoring has obvious significance for the early prediction and accurate diagnosis of cardiovascular diseases.However,in the remote ECG monitoring system,the interference of the external environment will cause the dynamic ECG signal to contain a large amount of noise,which will affect the detection of the ECG signal waveform,resulting in misjudgement of the intelligent analysis of the remote ECG monitoring system.Therefore,the problem of dynamic ECG signal de-noising has gradually become the focus and difficulty in the research of cardiovascular diseases in the context of telemedicine.Under this background,this paper combines factor analysis with gradient descent algorithm to study the noise reduction of dynamic ECG signal.The main research contents of the paper are as follows:(1)An optimization method with supervision factor analysis is proposed aiming at the optimization of depth factor analysis.Firstly,the optimization objective function is designed,and the partial derivative of the optimization objective function to the parameters is deduced.Then use the gradient descent algorithm for supervised parameter tuning training to solve the optimal parameters.(2)An ECG signal de-noising algorithm based on supervised learning depth factor analysis was proposed aiming at the de-noising problem of ECG signal.Using factor analysis to construct depth factor analysis model of noisy ECG signal,the signal and noise can be distinguished from each other.And the supervised factor analysis optimization method can be used to reconstruct the network at the top of the network.The top-level factor load matrix is trained with supervision to obtain the optimal factor load matrix,and then the ECG signal is reconstructed.More noise can be filtered from ECG signals because there is a supervised training process to make full use of the features of the noise,so that better noise reduction effect can be achieved.(3)The proposed ECG signal de-noising algorithm is applied to a smart ECG monitoring platform.In order to verify the practicability and effectiveness of this paper,the ECG signal de-noising algorithm for supervised learning depth factor analysis was applied to the intelligent ECG monitoring platform independently researched and developed by the laboratory task force.It is shown that the proposed algorithm can effectively filter the noise of ECG signals while maintaining the main characteristics of ECG signals by the actual electrocardiogram information passed to users randomly and anonymously.