Research On Intelligent Detection And Localization Of Myocardial Infarction Based On Machine Learning

Myocardial infarction(MI)is an acute ischemic heart disease with extremely high mortality,which is prone to malignant arrhythmia,heart rate failure,cardiogenic shock and other complications.According to China Cardiovascular Health Disease Report Writing Group,the domestic cardiovascular mortality rate in the stages of China is still hosted first,higher than tumor and other diseases.The most important cause of death in cardiovascular disease is MI,and when MI occurs,the patient has a valuable gold treatment time.If the patients are diagnosed in time and obtain the corresponding treatment,the area of MI can be effectively controlled,and the mortality is reduced,so early detection and positioning research on MI is important.The electrocardiogram(ECG)is one of the early detection methods of cardiovascular disease.This thesis is aimed at the presence of the model training of the existing MI and the problem of positioning accuracy does not meet clinical diagnosis,and the time domain,frequency domain characteristics,and the pilot between the electrocardiographic signals are combined.Correlation information between heart rate cycle to achieve early diagnosis and positioning of MI,and the specific research content are as follows:Aiming at the problem of early detection of MI and the ECG signals with a time of 10 s collected by existing ECG devices,this thesis fuses the characteristics of the time domain and frequency domain in the lead of ECG signals.By specific changes in the ECG caused by the ECG sed by MI,extraction of 12-lead electrical signals of each lead of QRS wave and ST-T wave amplitude and area and S wave of ECG heart beat most value point and T waves the value point of the slope,fusion ECG signal by wavelet transform in the frequency domain signal after the sample entropy.For the existing MI,six classifiers are divided,and the genetic algorithm is screened by genetic algorithm,screening,respectively,the characteristics of the sensitivity of MI in different sites,thereby achieving early rapid detection of MI,and provide guidance effects for the judgment of clinical center muscle infarction.The experiment proved that the accuracy,sensitivity and specificity of MI detection reached 98.69%,99.22% and98.14% respectively.This paper is directed to the PEARSON correlation coefficient between the PEARSON related coefficient between the different leads of the 12 lead-connected card electrical signal and the PEARSON correlation coefficient between each heart rate cycle in the pilot data.The evolution characteristics of the center rate cycle of MI and the coupling characteristics between the leaders.The time domain and frequency domain signal in the leader are then fused,and the expression characteristics in the electrocardiographic signal are fully extracted,and the early precision localization of MI is achieved.Experiments have proven that in the localization of MI at different parts,the average accuracy,sensitivity,and specificity reached99.94%,99.77%,and 99.93%.