CN108960113A - A kind of heart rate variability recognition methods based on support vector machines - Google Patents

A kind of heart rate variability recognition methods based on support vector machines Download PDF

Info

Publication number
CN108960113A
CN108960113A CN201810671562.7A CN201810671562A CN108960113A CN 108960113 A CN108960113 A CN 108960113A CN 201810671562 A CN201810671562 A CN 201810671562A CN 108960113 A CN108960113 A CN 108960113A
Authority
CN
China
Prior art keywords
support vector
sample
vector machines
heart rate
rate variability
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201810671562.7A
Other languages
Chinese (zh)
Inventor
岳大超
刘海宽
张磊
李致远
蒋大伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu Normal University
Original Assignee
Jiangsu Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jiangsu Normal University filed Critical Jiangsu Normal University
Priority to CN201810671562.7A priority Critical patent/CN108960113A/en
Publication of CN108960113A publication Critical patent/CN108960113A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Signal Processing (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

The heart rate variability recognition methods based on support vector machines that the invention discloses a kind of, belongs to intelligent medical technical field.In the heart rate variability recognition methods, firstly, obtaining electrocardiogram (ECG) data, and the data obtained is handled, extract R crest value position；Secondly, calculating heartbeat interval, RdR scatter plot is drawn accordingly, rear sample drawn is normalized to the figure and is marked, and sample is sampled, randomly select 75% as training sample, remaining is as test sample；Again, learnt with support vector machines, then tested with test sample.The present invention provides basis to realize automated diagnostic, the waste alleviated medical resource in short supply, reduce medical resource, improving medical efficiency.

Description

A kind of heart rate variability recognition methods based on support vector machines
Technical field
The present invention relates to intelligent medical technical field, in particular to a kind of heart rate variability based on support vector machines is known Other method.
Background technique
Heart rate variability refers to the time-variance number of cardiac cycle, and research object is cardiac cycle rather than heart rate.People's Heart rate be not it is unalterable, there is small time differences between heartbeat twice, calculate the difference of cardiac cycle Understand heart rate variability (Heart rate variability, HRV).Heart rate variability can assess cardiac sympathetic nerve with The bulk information of cardiovascular aspect is contained in influence of the parasympathetic nerve to cardiovascular activity.Clinical research shows that heart rate becomes Anisotropic reduction is the symptom of the cardiovascular disease incidences such as myocardial infarction, hypertension, angina pectoris.Therefore, heart rate variability is ground Study carefully, has in the breaking-out evaluated systema cariovasculare functional, predict cardiovascular disease, and for the early diagnosis of cardiovascular disease Important meaning.
Poincare scatter plot is a kind of important research method of heart rate variability.By using continuous heartbeat interval The graphing in rectangular coordinate system reflects the variation of adjacent heartbeat interval, can show the feature of heartbeat interval.Poincare For scatter plot there are many form, including comet formation, sector etc., different shapes reflects different heart states.Although Poincare Scatter plot is a kind of effective heart rate variance analyzing method, but can not embody its trend changed over time, to Mr. Yu A little cardiovascular disease, physical conditions etc. cannot embody its heart rate variability property well.Then, some scholars propose Improvement strategy, i.e. first order difference plot, by the difference of adjacent heartbeat interval draw scatter plot.However, this method It is lost the absolute value information of original heartbeat interval again.Therefore, there is scholar to combine the two, it is scattered to propose a kind of RdR Point diagram is come with this while reflecting heartbeat interval and its variation.
Support vector machines (Support Vector Machine, SVM) is that Vapnik etc. develops on the basis of SLT, It is a kind of to study very widely used machine learning algorithm.Support vector machines is in the case where fixed empiric risk, by most Structural risk minimization is realized at bigization edge, so that learning machine obtains satisfied learning effect, has very strong generalization ability.
The application is to carry out identification sort research to RdR scatter plot by support vector machines.
Summary of the invention
The classifier that it is an object of the invention to be obtained by using support vector machines carrys out automatic identification RdR scatter plot, It performs an analysis to heart rate variability, to realize automated diagnostic, the waste alleviated medical resource in short supply, reduce medical resource, mentioning The medical efficiency of height provides basis.
To solve the above-mentioned problems, the technical scheme is that
1, a kind of heart rate variability recognition methods based on support vector machines, which comprises the following steps:
Step 1) obtains electrocardiosignal, and is filtered denoising to the electrocardiosignal, extracts R crest value position；
Step 2) draws RdR scatter plot using heartbeat interval；
Step 3) zooms in and out the RdR scatter plot, changes into grayscale image, and image data is normalized, To reduce calculation amount；
The sample in scatter plot that step 4) obtains step 3) is marked；
Step 5) samples the sample marked in step 4), randomly selects 75% data as training sample, Remaining is as test sample；
Step 6) carries out dimension-reduction treatment to data using principal component analysis, reduces redundancy feature；
Support vector machines parameter is arranged in step 7), is trained with the support vector machines by the training sample, is Support vector machines overlearning is prevented, penalty coefficient is introduced；
Step 8) was tested originally with the test specimens, to assess the performance for using the resulting classifier of above-mentioned steps, if full Sufficient performance indicator then obtains disaggregated model, otherwise return step 7) Reparametrization is trained.
Compared with the existing technology, the beneficial effects of the present invention are:
The present invention is to realize automated diagnostic, the waste alleviated medical resource in short supply, reduce medical resource, improve just It examines efficiency and basis is provided.
Detailed description of the invention
Fig. 1 is that the present invention is based on the flow charts of the heart rate variability recognition methods of support vector machines；
Fig. 2 is the Partial Feature figure that the application is extracted from RdR scatter plot；
Fig. 3 is to use the resulting part of test results figure of heart rate variability recognition methods described herein.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts it is all its His embodiment, shall fall within the protection scope of the present invention.
The present embodiment can be selected in Intel Xeon CPU E5-2697 2.70GHz, inside save as 128.00GB, Win7, It is realized in the computer of 64 bit manipulation systems.The present embodiment is made schematic with the data of the electrocardiogram (ECG) data in MIT-BIH database Analysis introduction.The support vector machines of identification heart rate variability can be obtained through recognition methods provided in this embodiment, that is, obtains one The classifier of the recognizable heart rate variability of kind.
Please refer to Fig. 1-3, a kind of heart rate variability recognition methods based on support vector machines, comprising the following steps:
Step 1) acquires electrocardiosignal, to obtain electrocardiogram (ECG) data；Denoising is filtered to electrocardiosignal, extracts R wave Peak position；
Step 2) calculates heartbeat interval, draws RdR scatter plot using heartbeat interval, drawing tool can be MATLAB etc.； RdR scatter plot is a kind of heart rate variance analyzing method, can embody heart rate variability and change with time trend；
Step 3) handles image data, specifically: zooming in and out to RdR scatter plot, change into grayscale image, and to image Data are normalized, to reduce calculation amount；
Step 4) extraction step 3) it sample in obtained scatter plot and is marked；
Step 5) samples the sample marked in step 4), randomly selects 75% data as training sample, Remaining is as test sample；
Step 6) carries out dimension-reduction treatment to data using principal component analysis (PCA), reduces redundancy feature；
Support vector machines parameter is arranged in step 7), is trained with support vector machines, to prevent support vector machines from excessively learning It practises, introduces penalty coefficient；In order to make it easy to understand, the present embodiment illustrates Partial Feature by Fig. 2: RdR scatter plot passes through data After normalization, PCA dimension-reduction treatment, feature is extracted using SVM (support vector machines) and is obtained, Partial Feature figure is as shown in Figure 2.
Step 8) the original test performance of test specimens, to assess the performance of the resulting classifier of the present invention, if meeting performance Index then obtains sorter model, otherwise return step 7) Reparametrization is trained.
The classification results of method according to the present invention, the classifier of acquisition are as shown in Figure 3, wherein parameter true is practical Sample class, predicted be classifier prediction classification.
Specifically, support vector machines is a kind of machine learning algorithm, maximized by edge, realizes Structural risk minization Change, to obtain optimal classifying quality.The basic skills of support vector machines are as follows:
Given training sample D={ (x1,y1),(x2,y2),…,(xm,ym)},yi∈ { -1 ,+1 }, wherein x1,x2……xm That indicate is sample data, y1,y2……ymWhat is indicated is sample class；The classificating thought of support vector machines is in sample sky Between in find a hyperplane, sample can be distinguished.In sample space, Optimal Separating Hyperplane can pass through such as lower section Journey indicates:
ωTX+b=0
Wherein, ω=(ω1；ω2；...；ωd) it is normal vector, b is displacement item.
If hyperplane can correctly classify sample, i.e., for (xi,yi) ∈ D, if yi=+1, there is ωTxi+ b > 0；If yi , there is ω in=- 1Txi+ b < 0.Then enable
The sample nearest apart from hyperplane sets up the equal sign of above formula, is supporting vector, two inhomogeneous supports The sum of the distance of vector and hyperplane isReferred to as it is spaced.The hyperplane for finding largest interval is sought under the constraints, So that margin maximization, it may be assumed that
s.t.yi(ωTxi+ b) >=1, i=1,2 ..., m
Wherein, s.t. indicates constraint condition；Here it is the basic models of support vector machines.
However, sample is not linear separability in most of situations in realistic task, it in this case, can Sample to be mapped to the space of more higher-dimension, sample is linear separability in this higher dimensional space.
Enable φ (x) indicate the feature vector after mapping x, i.e., by some function phi (x), by x from former spatial alternation to Another space, then Optimal Separating Hyperplane is represented by
F (x)=ωTφ(x)+b
The basic model formula of support vector machines can be rewritten as
s.t.yi(ωTφ(xi)+b) >=1, i=1,2 ..., m
Its dual problem is
In above formula, φ (xi)Tφ(xj) it is inner product of the sample in higher dimensional space, αiWith αjIt is all Lagrange multiplier；Due to Space dimensionality may be very high, more difficulty is directly calculated, it is therefore contemplated that just like minor function
k(xi,xj)=< φ (xi),φ(xj) >=φ (xi)Tφ(xj)
Bring Optimal Separating Hyperplane formula into, after solution
Wherein, k (x, xi) it is kernel function.Calculating can greatly be simplified by kernel function, common kernel function is wired Property core, Gaussian kernel, polynomial kernel etc..
In actual assorting process, even if making sample linear separability by kernel function, it is also difficult to which guarantee is not overfitting Caused by, to alleviate this problem, the concept of soft margin is introduced, that is, certain samples is allowed to be unsatisfactory for constraint condition.Optimization Target can be rewritten as
s.t.yi(ωTxi+b)≥1-ξi, i=1,2 ..., m
Wherein, C is constant, ξiIt is slack variable, here it is soft margin support vector machines.Each sample corresponds to One slack variable, to characterize the degree that sample is unsatisfactory for constraint.
Specifically, the basic skills of principal component analysis (PCA) are as follows:
Before classifying to picture, in order to reduce redundancy feature and calculation amount, it is necessary first to be carried out to characteristics of image Dimensionality reduction, the present embodiment extract main feature using principal component analysis (Principal Component Analysis, PCA), Dimension-reduction treatment is carried out to image data.
It is assumed that having carried out the new coordinate system after centralization processing and projective transformation to sample is { a1,a2,...,ad, wherein aiIt is normal orthogonal base vector, | | ai||2=1,If abandoning the partial coordinates in new coordinate system, dimension drop For d'< d, then sample is projected as z in the coordinate system of low-dimensionali=(zi1；zi2；...；zid'),It is xiIt is sat in low-dimensional Mark is that lower jth ties up coordinate.Based on ziTo reconstruct xi, obtainConsider entire training sample, original sample xiWith base Sample after reconstructThe distance between be
Wherein, W=(a1,a2,...,ad).According to recently it is reconstitution, the optimization aim of principal component analysis is
s.t.WTW=I
Wherein, X=(x1,x2,...,xm), dimension d' after dimensionality reduction is usually specified in advance, can also pass through setting Threshold value is reconstructed, such as t=95%, then choosing makesThe d' minimum value of establishment is dimension values.
The application is illustrated in conjunction with specific embodiments above, however, this application is not limited to this, any this field Technical staff can think variation, should all fall in the protection domain of the application.

Claims (1)

1. a kind of heart rate variability recognition methods based on support vector machines, which comprises the following steps:
Step 1) obtains electrocardiosignal, and is filtered denoising to the electrocardiosignal, extracts R crest value position；
Step 2) draws RdR scatter plot using heartbeat interval；
Step 3) zooms in and out the RdR scatter plot, grayscale image is changed into, and image data is normalized, to subtract Few calculation amount；
The sample in scatter plot that step 4) obtains step 3) is marked；
Step 5) samples the sample marked in step 4), randomly selects 75% data as training sample, remaining It is test sample；
Step 6) carries out dimension-reduction treatment to data using principal component analysis, reduces redundancy feature；
Support vector machines parameter is arranged in step 7), is trained with the support vector machines by the training sample, to prevent The support vector machines overlearning introduces penalty coefficient；
Step 8) was tested originally with the test specimens, to assess the performance for using the resulting classifier of above-mentioned steps, if satisfaction property Can index then obtain disaggregated model, otherwise return step 7) Reparametrization is trained.
CN201810671562.7A 2018-06-26 2018-06-26 A kind of heart rate variability recognition methods based on support vector machines Pending CN108960113A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810671562.7A CN108960113A (en) 2018-06-26 2018-06-26 A kind of heart rate variability recognition methods based on support vector machines

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810671562.7A CN108960113A (en) 2018-06-26 2018-06-26 A kind of heart rate variability recognition methods based on support vector machines

Publications (1)

Publication Number Publication Date
CN108960113A true CN108960113A (en) 2018-12-07

Family

ID=64486898

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810671562.7A Pending CN108960113A (en) 2018-06-26 2018-06-26 A kind of heart rate variability recognition methods based on support vector machines

Country Status (1)

Country Link
CN (1) CN108960113A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110910021A (en) * 2019-11-26 2020-03-24 上海华力集成电路制造有限公司 Method for monitoring online defects based on support vector machine
CN111839494A (en) * 2020-09-04 2020-10-30 广东电网有限责任公司电力科学研究院 Heart rate monitoring method and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101923648A (en) * 2009-06-15 2010-12-22 深圳迈瑞生物医疗电子股份有限公司 Clustering method and device for support vector machine
CN103584840A (en) * 2013-11-25 2014-02-19 天津大学 Automatic sleep stage method based on electroencephalogram, heart rate variability and coherence between electroencephalogram and heart rate variability
CN104127194A (en) * 2014-07-14 2014-11-05 华南理工大学 Depression evaluating system and method based on heart rate variability analytical method
CN105496402A (en) * 2015-11-20 2016-04-20 北京理工大学 Electrocardio feature analyzing method based on point diagram and symbolic dynamics

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101923648A (en) * 2009-06-15 2010-12-22 深圳迈瑞生物医疗电子股份有限公司 Clustering method and device for support vector machine
CN103584840A (en) * 2013-11-25 2014-02-19 天津大学 Automatic sleep stage method based on electroencephalogram, heart rate variability and coherence between electroencephalogram and heart rate variability
CN104127194A (en) * 2014-07-14 2014-11-05 华南理工大学 Depression evaluating system and method based on heart rate variability analytical method
CN105496402A (en) * 2015-11-20 2016-04-20 北京理工大学 Electrocardio feature analyzing method based on point diagram and symbolic dynamics

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
李旭超等: "图像分割中的模糊聚类方法", 《中国图象图形学报》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110910021A (en) * 2019-11-26 2020-03-24 上海华力集成电路制造有限公司 Method for monitoring online defects based on support vector machine
CN111839494A (en) * 2020-09-04 2020-10-30 广东电网有限责任公司电力科学研究院 Heart rate monitoring method and system

Similar Documents

Publication Publication Date Title
Hussain Detecting epileptic seizure with different feature extracting strategies using robust machine learning classification techniques by applying advance parameter optimization approach
Moghaddasi et al. Automatic assessment of mitral regurgitation severity based on extensive textural features on 2D echocardiography videos
Jin et al. A novel domain adaptive residual network for automatic atrial fibrillation detection
Yu et al. An automatic computer-aided detection scheme for pneumoconiosis on digital chest radiographs
Gudigar et al. Global weighted LBP based entropy features for the assessment of pulmonary hypertension
CN111046731B (en) Transfer learning method and recognition method for gesture recognition based on surface electromyographic signals
Bhuvaneswari et al. A new fusion model for classification of the lung diseases using genetic algorithm
Bhuiyan et al. Transfer learning and supervised classifier based prediction model for breast cancer
Suresha et al. Alzheimer disease detection based on deep neural network with rectified Adam optimization technique using MRI analysis
CN109817339A (en) Patient&#39;s group technology and device based on big data
CN112598661A (en) Ankle fracture and ligament injury diagnosis method based on machine learning
Islam et al. Analysis of CT scan images to predict lung cancer stages using image processing techniques
CN109284781A (en) Image classification algorithms and system based on manifold learning
Mohan et al. Fast and robust exudate detection in retinal fundus images using extreme learning machine autoencoders and modified kaze features
Hussain et al. Deep learning-based diagnosis of disc degenerative diseases using MRI: a comprehensive review
Zhou et al. Automatic microaneurysms detection based on multifeature fusion dictionary learning
Prakash et al. A system for automatic cardiac arrhythmia recognition using electrocardiogram signal
Liu et al. Automatic detection and quantification of brain midline shift using anatomical marker model
Zhang et al. Detection and quantification of intracerebral and intraventricular hemorrhage from computed tomography images with adaptive thresholding and case-based reasoning
Bagaria et al. Bone fractures detection using support vector machine and error backpropagation neural network
Carvalho et al. Method of differentiation of benign and malignant masses in digital mammograms using texture analysis based on phylogenetic diversity
CN108960113A (en) A kind of heart rate variability recognition methods based on support vector machines
Bansal et al. An improved hybrid classification of brain tumor MRI images based on conglomeration feature extraction techniques
Kapila et al. Efficient feature selection technique for brain tumor classification utilizing hybrid fruit fly based abc and ann algorithm
US20230371872A1 (en) Method and system for quantifying attention

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20181207

WD01 Invention patent application deemed withdrawn after publication