阅读说明：本技术 基于连续心搏活动序列特征-svm模型的心搏分类方法 (Heart beat classification method based on continuous heart beat activity sequence feature-SVM model ) 是由 李润川 申圣亚 张宏坡 陈刚 王宗敏 于 2020-05-22 设计创作，主要内容包括：本发明涉及一种基于连续心搏活动序列特征-SVM模型的心搏分类方法,包括以下步骤：S1、采用连续小波变化去除心电信号中的噪声；S2、对经过步骤S1处理后的心电信号进行分割截取出完整的心搏,然后从截取出的心搏中进行特征提取,并将提取到的特征按类别建立以下数据集：集合A＝{470单心搏形态特征},集合B＝{21个连续RR间期}集合C＝{491连续心搏全局序列特征}；S3、将步骤S2中的数据集中的任一个集合输入到采用SVM算法模型中进行心搏分类；本发明具有良好的心搏分类准确性。(The invention relates to a heart beat classification method based on a continuous heart beat activity sequence feature-SVM model, which comprises the following steps: s1, removing noise in the electrocardiosignals by adopting continuous wavelet change; s2, segmenting the electrocardiosignal processed in the step S1 to obtain complete heart beats, extracting features from the heart beats, and establishing the following data sets according to the extracted features: a set A is {470 single heart beat morphological characteristics }, a set B is {21 continuous RR intervals } and a set C is 491 continuous heart beat global sequence characteristics }; s3, inputting any set of the data sets in the step S2 into an SVM algorithm model for heart beat classification; the invention has good heart beat classification accuracy.)

1. A heart beat classification method based on a continuous heart beat activity sequence feature-SVM model is characterized by comprising the following steps:

s1, removing noise in the electrocardiosignals by adopting continuous wavelet change;

s2, segmenting the electrocardiosignal processed in the step S1 to obtain complete heart beats, extracting features from the heart beats, and establishing the following data sets according to the extracted features:

the set a ═ 470 single-heartbeat morphology features },

set B ═ 21 consecutive RR intervals },

set C ═ 491 continuous heart beat global sequence feature };

s3, inputting any one of the data sets in step S2 into a SVM algorithm model for heart beat classification.

2. The heart beat classification method based on continuous heart beat activity sequence feature-SVM model according to claim 1, characterized in that: the 470 single-heart-beat morphological characteristics extraction method is that 235 sampling points near the R peak are respectively extracted through two leads contained in each record of the MIT-BIH arrhythmia database by utilizing the R peak position determined in the annotation file.

Technical Field

The invention belongs to the technical field of arrhythmia classification methods, and discloses a heart beat classification method based on a continuous heart beat activity sequence feature-SVM model.

Background

The normal heart has four physiological functions: autonomic, excitability, conductibility and contractility. Arrhythmia refers to a cardiac arrhythmia or a frequency of heart beats caused by an error or disorder in the pacing position or conduction of the heart impulses. Aiming at the continuous change of the heart beat waveform, the computer can accurately and effectively read the electrocardiogram and gradually give out the diagnosis result. At present, an intelligent classification mode is gradually adopted in the electrocardiographic diagnosis, for example, the current GaussianNB (gaussian bayes), linear discriminant analysis (linear discriminant analysis), logistic regression (logistic regression), decision tree (decision tree), GBDT (gradient boosting iterative decision tree), RandomForest (random forest), AdaBoost (adaptive enhancement) is applied to intelligent cardiac beat classification, which has a good effect of improving the efficiency and accuracy of cardiac beat classification in the intelligent electrocardiographic classification.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provide a heart beat classification method based on a continuous heart beat activity sequence feature-SVM model.

The technical scheme of the invention is as follows:

a heart beat classification method based on a continuous heart beat activity sequence feature-SVM model comprises the following steps:

s1, removing noise in the electrocardiosignals by adopting continuous wavelet change;

s2, segmenting the electrocardiosignal processed in the step S1 to obtain complete heart beats, extracting features from the heart beats, and establishing the following data sets according to the extracted features:

the set a ═ 470 single-heartbeat morphology features },

set B ═ 21 consecutive RR intervals },

set C ═ 491 continuous heart beat global sequence feature };

s3, inputting any one of the data sets in step S2 into a SVM algorithm model for heart beat classification.

Further, the 470 single-heart-beat morphological characteristics extraction method is that 235 sampling points near the R peak are respectively extracted through two leads contained in each record of the MIT-BIH arrhythmia database by using the R peak position determined in the annotation file.

Compared with the prior art, the invention has the beneficial effects that:

according to the method, the noise is removed through continuous wavelet transformation, then the characteristic extraction is carried out, three sets of the extracted characteristics are established according to the types of the characteristics, one of the three sets is input into an SVM (support vector machine) model for heart beat classification, the set A, the set B and the set C are used as the input of the electrocardio classification, the SVM model is used as a classifier for classifying the electrocardio signals in the sets, and the accuracy of the electrocardio classification can be effectively improved.

Drawings

FIG. 1 is a schematic illustration of an optimal separation hyperplane.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

A heart beat classification method based on a continuous heart beat activity sequence feature-SVM model comprises the following steps:

s1, removing noise in the electrocardiosignals by adopting continuous wavelet change;

s2, segmenting the electrocardiosignal processed in the step S1 to obtain complete heart beats, extracting features from the heart beats, and establishing the following data sets according to the extracted features:

the set a ═ 470 single-heartbeat morphology features },

set B ═ 21 consecutive RR intervals },

set C ═ 491 continuous heart beat global sequence feature };

s3, inputting any one of the data sets in step S2 into a SVM algorithm model for heart beat classification.

Further, the 470 single-heart-beat morphological characteristics extraction method is that 235 sampling points near the R peak are respectively extracted through two leads contained in each record of the MIT-BIH arrhythmia database by using the R peak position determined in the annotation file.

The SVM model adopted in the invention is a novel machine learning method established on the basis of a statistic learning theory VC dimension theory and a structure risk minimization principle, the algorithm shows a plurality of specific advantages in solving the problems of small sample, nonlinearity and high dimension pattern recognition, and the problems of dimension disaster, over-learning and the like are overcome to a great extent, the basic idea of the SVM is shown in figure 1, hollow points and solid points respectively represent different categories, H is a separating hyperplane, and H is a separating hyperplane₁And H₂Is called a support vector, H₁And H₂Referred to as classification interval; the optimal separation hyperplane is the one requiring the largest classification interval on the premise of correctly separating different classes.

Assume that the linear classification plane is of the form:

g(x)＝w^Tx+b (1)；

wherein w is a classification weight vector, b is a classification threshold, the discriminant function is normalized so that the discriminant function satisfies | g (x) | ≧ 1 for both types of samples,

y_i(w^Tx_i+b)-1≥0，i＝1，2，…，l (2)；

(2)

wherein y is_iIs a class label of the sample, x_iAre the corresponding samples.

The maximization of the classification interval Margin 2/(| w |) is equivalent to the calculation ofAnd minimum.

Introducing lagrange multiplier a_iAccording to the Karush-Kuhn-Tucker (KKT) condition, the above problem can be converted into the maximization of the generic function w (a) under the constraint condition (4), and the generalizationThe expression of the function w (a) is shown in (5).

∑^l _i＝1α_iy_i＝0，α_i≥0，i＝1，2，…，l (4)；

w(α)＝∑^l _i＝1α_i-1/2 ∑^l _i＝1∑^l _j＝1α_iα_jy_iy_j＜x^T _ix_j＞ (5)；

Quadratic programming can find a_iA is to_iSubstituting into (6) to obtain w;

w＝∑^l _i＝1α_iy_jx_x(6)；

selecting a not zero_iSubstituting the result into the step (7) to obtain b;

α_i(y_j(w^Tx_i+b)-1)＝0 (7)：

by derivation, the decision function becomes the following equation:

f(x)＝sign(∑_svy_jα_i＜x^T _ix_j＞+b) (8)；

the test sample is substituted into formula (8), and if f (x) is 1, the test sample belongs to the category, otherwise, the test sample does not belong to the category.

Experiments and results are as follows:

the following experiments were all performed on the MIT-BIH arrhythmia database, and each heart beat was classified as N (normal or bundle branch block), S (supraventricular abnormal heart beat), V (ventricular abnormal heart beat), F (fused heart beat), Q (unclassified heart beat) according to ANSI/AAMI EC 57.

In this embodiment, four values of TP, FP, TN and FN are calculated to obtain the heart beat classification result, where TP is used for classifying heart beats_NRepresenting the N-class of true-positive heartbeats, FP_NRepresenting false positive heartbeats of type N, TN_NN true negative heart beats are shown, and FNN false negative heart beats are shown. The classification results for other heart beat classes are calculated in the same way. Table 1 shows the confusion matrix of the classification results. In the experiment of the embodiment, N, S, V, F and Q represent the real categories of heartbeats; n, s, v, f, q represent the prediction results。

Table 1: confusion matrix of classification results

TP_N＝N_N(9)

FP_N＝Ns+Nv+Nf+Nq (10)

TN_N＝Ss+Sv+Sf+Sq+Vs+Vv+Vf+Vq+Fs+Fv+Ff+Fq+Qs+Qv+Qf+Qq (11)

FN_N＝Sn+Vn+Fn+Qn (12)

This implementation uses sensitivity, specificity, positive predictive value, and accuracy to evaluate classifier performance. Sensitivity (se) refers to the proportion of samples judged to be positive examples to all positive examples. The higher the sensitivity, the more samples are correctly predicted. Specificity (sp) refers to the proportion of samples judged to be negative in all negative cases. The positive predictive value (+ p) is also referred to in the literature as accuracy. Accuracy is the ratio of the number of correctly classified samples to the total number of samples, reflecting the consistency between the test results and the actual results. The calculation formulas of the four evaluation indexes are as follows:

Se＝TP/(TP+FN)(13)

Sp＝TN/(TN+FP) (14)

+p＝TP/(TP+FP) (15)

Acc＝(TP+TN)/(TP+TN+FP+FN) (16)。

for analysis of experimental results using different combinations of features of one or more of the foregoing components of set A, B, C, D, E, the analysis proceeds as follows:

experiment I

Experiment I is based on the set A, the set B and the set C as input, GaussianNB (Gaussia Bayes), Linear discriminant analysis (linear discriminant analysis), Logistic regression (logistic regression), DecisionTree (decision tree), GBDT (gradient boosting iterative decision tree), RandomForest (random forest), AdaBoost (adaptive boosting) as reference classifiers, SVM models as experiment classifiers are used for heart beat classification analysis, and classification results of different model classifiers based on different data sets are shown in the following table 2.

TABLE 2 Classification results for different models based on different datasets

Experiment II

Experiment II heart beat classification based on set a and SVM model. The experimental results show that the average classification accuracy of the heart beat classification is 98.96%, but only the morphology features of the single heart beat are used, so that the method has certain locality. The disadvantage of this experimental method is that the morphological characteristics of single heart beat are too unilateral. Table 3 shows the classification performance of the SVM model with only 470 single heart beat morphology features.

TABLE 3 Classification results and Performance based on set A and SVM models

Experiment III

Experiment III cardiac beat classification based on set B and SVM models. The experimental results show that the classification precision of N, S, V and Q classes is obviously low. The average classification accuracy reached 88.46%. Table 4 shows the classification results for SVM models featuring only continuous inter-beat activity. . The disadvantage of this experimental approach is that the continuous inter-beat activity profile is also not comprehensive enough.

TABLE 4 Classification results and Performance based on set B and SVM models

Experiment IV

Experiment IV cardiac beat classification based on set C and SVM model. The experimental results show that the average accuracy of classification reaches 99.12%. The overall accuracy of experiment IV was improved by 0.16% compared to experiment II. The overall accuracy was improved by 10.66% compared to experiment III. Table 5 shows the classification results and performance of Adaboost + random forest models with QRS area features only.

TABLE 5 Classification results and Performance based on set D and sVM models

In summary, according to the heart beat classification method based on the continuous heart beat activity sequence feature-SVM model provided by the application, a set a ═ 470 single heart beat morphology feature }, a set B ═ 21 continuous RR intervals, and a set C ═ 491 continuous heart beat global sequence feature } are used as input, the SVM model is used as a classifier to classify heart beats of an electrocardiosignal, the classification accuracy can reach 99.12%, and the classification accuracy is higher for the input electrocardiosignal data set compared with that of a traditional classifier.

Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that various changes in the embodiments and/or modifications of the invention can be made, and equivalents and modifications of some of the features described in the embodiments and/or modifications can be made without departing from the spirit and scope of the invention.