The Research On J Wave Detection Techniques Based On Decision Tree

J wave is potential changes of the QRS wave end point and ST segment starting point in the electrocardiogram(ECG).J wave can cause a series of syndromes and lead to heart disease such as malignant arrhythmia and sudden cardiac death.At present,the detection of J wave also mainly relies on the doctor's clinical experience,which will cause great error.Through the method of signal processing,the rapid and accurate identification and detection of J wave can prevent the occurrence of J wave syndrome.Based on the idea of multi-layer classification,this paper aims to detect and identify the J wave by Kalman filter and decision tree classification algorithm.Before the detection,the wavelet de-noising method is used to eliminate the noise of the ECG signal,and the influence of the power frequency noise,EMG noise and baseline drift on the J wave detection is eliminated.After denoising,the signal to noise ratio is obtained,which lays the foundation for the accurate detection of J wave.Then,the Kalman filter is used to detect ECG signal.The Kalman filter is used to carry out the early warning of abnormal ECG signal,and the collected data are screened,so as to reduce the false detection rate of J wave detection and recognition.After ECG pre-detection,the data set can only be left abnormal ECG records in the ST segment.In order to construct an effective J wave detection model,this paper carries out a multi-directional feature extraction of ECG signals and extracts six time-frequency features,including two types of time domain features and four types of frequency domain features.After that,correlation analysis and the dimension reduction of manifold algorithm are carried out for the extracted features,the purpose is to achieve the optimization of features,reduce the complexity of the algorithm and reduce the cost of computing.Then the decision tree is constructed by using the improved information gain algorithm,and the algorithm is stabilized by training.In order to improve the precision of the decision tree,this paper uses the improved sample selection method to train and test the decision tree.The results show that the improved algorithm reduces the error rate of J wave detection to a certain extent.The results show that the accuracy of the J wave detection method is96.16%,and the specificity and sensitivity are more than 95%.Moreover,the detection performance is better than other algorithms in the same field,which achieves a higher evaluation standard.In general,this method can effectively detect and identify J waves,which can provide some help for clinical prevention and diagnosis of J wave.