CN109119158A - A kind of RdR scatter plot recognition methods based on sparse kernel principle component analysis - Google Patents
A kind of RdR scatter plot recognition methods based on sparse kernel principle component analysis Download PDFInfo
- Publication number
- CN109119158A CN109119158A CN201810884233.0A CN201810884233A CN109119158A CN 109119158 A CN109119158 A CN 109119158A CN 201810884233 A CN201810884233 A CN 201810884233A CN 109119158 A CN109119158 A CN 109119158A
- Authority
- CN
- China
- Prior art keywords
- rdr
- component analysis
- scatter plot
- spe
- sample
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000004458 analytical method Methods 0.000 title claims abstract description 8
- 239000013598 vector Substances 0.000 claims abstract description 17
- 238000012360 testing method Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 8
- 238000010586 diagram Methods 0.000 claims description 25
- 238000000513 principal component analysis Methods 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims 1
- 239000000284 extract Substances 0.000 abstract 1
- 239000011159 matrix material Substances 0.000 description 9
- 208000024172 Cardiovascular disease Diseases 0.000 description 6
- 230000000747 cardiac effect Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012847 principal component analysis method Methods 0.000 description 2
- 206010002383 Angina Pectoris Diseases 0.000 description 1
- 206010020772 Hypertension Diseases 0.000 description 1
- 230000005792 cardiovascular activity Effects 0.000 description 1
- 210000000748 cardiovascular system Anatomy 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013399 early diagnosis Methods 0.000 description 1
- 230000002526 effect on cardiovascular system Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 208000010125 myocardial infarction Diseases 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 210000005037 parasympathetic nerve Anatomy 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000002889 sympathetic effect Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/15—Biometric patterns based on physiological signals, e.g. heartbeat, blood flow
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Biomedical Technology (AREA)
- Multimedia (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Pathology (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The present invention discloses a kind of RdR scatter plot recognition methods based on sparse kernel principle component analysis, including step 1) obtains electrocardiosignal, is filtered denoising to electrocardiosignal, extracts R crest value position；Step 2 draws RdR scatter plot using heartbeat interval；Step 3) zooms in and out RdR scatter plot, and image data is normalized；The scatter plot sample of acquisition is marked in step 4)；Step 5) samples sample；Parameter is arranged in step 6)；Step 7) solves approximate base, characteristic value, feature vector to every class sample respectively；Step 8) calculates the SPE of each test sample, compared with concrete class, the step 9 if meeting the requirements, and otherwise 6 Reparametrization of return step training；Step 9) obtains parameter, and algorithm terminates.The present invention carries out the Classification and Identification of RdR scatter plot using sparse kernel principle component analysis method, has satisfactory classifying quality.
Description
Technical Field
The invention relates to an RdR scatter diagram identification method based on sparse kernel principal component analysis, and belongs to the technical field of intelligent medical treatment.
Background
Heart rate variability refers to the time variation between cardiac intervals, which is the subject of the study. The Heart Rate Variability (HRV) can be understood by calculating the difference between cardiac intervals, which is not a constant human Heart rate, but a small time difference between two Heart beats. The heart rate variability allows assessment of the effect of the sympathetic and parasympathetic nerves on cardiovascular activity, containing a great deal of information on the cardiovascular aspect. Clinical studies indicate that the reduction in heart rate variability is a symptom of cardiovascular diseases such as myocardial infarction, hypertension, angina pectoris, and the like. Therefore, the study of heart rate variability is of great significance in evaluating cardiovascular system function, predicting the onset of cardiovascular disease, and for early diagnosis of cardiovascular disease.
The Poincare scattergram is an important research method for heart rate variability. By plotting the graph in the rectangular coordinate system using consecutive heart intervals, reflecting the variation of adjacent heart intervals, the characteristics of the heart intervals can be displayed. Poincare scattergrams have a variety of morphologies, including comet, fan, etc., with different shapes reflecting different cardiac states. Although the Poincare scatter diagram is an effective heart rate variability analysis method, the Poincare scatter diagram cannot reflect the change trend of the Poincare scatter diagram over time, and the heart rate variability of the Poincare scatter diagram cannot be well reflected for certain cardiovascular diseases, physical health conditions and the like. Thus, some scholars propose an improved strategy, namely a first order difference scatter plot, which is drawn by the difference of adjacent inter-cardiac intervals. However, this method loses the original absolute value information of the inter-beat intervals. Therefore, the combination of both was suggested by the scholars to provide an RdR scattergram to reflect both the inter-beat intervals and their changes.
Currently, heart rate variability is analyzed in many different cardiovascular diseases. However, how to identify and distinguish different cardiovascular diseases according to the scatter diagram is relatively poor. The RdR scatter diagram is identified and classified by a sparse kernel principal component analysis method.
Disclosure of Invention
The invention aims to automatically identify RdR scatter diagrams and analyze heart rate variability by a sparse kernel principal component analysis method, and provides a basis for realizing automatic diagnosis, relieving scarce medical resources, reducing the waste of medical resources and improving the diagnosis efficiency.
The invention realizes the purpose through the following technical scheme: a RdR scatter diagram identification method based on sparse kernel principal component analysis comprises the following steps:
step 1) acquiring electrocardiosignals, carrying out filtering and denoising treatment on the electrocardiosignals, and extracting the position of an R wave peak value;
step 2) drawing RdR scatter diagram by MATLAB tool;
step 3) scaling RdR scatter diagram, converting into grey map, and normalizing image data to reduce calculated amount;
step 4), marking the obtained scatter diagram sample;
step 5) sampling samples, and randomly extracting 75% of data as training samples;
step 6) setting parameters, wherein the parameters comprise error parameters of the approximate basis, Gaussian kernel function parameters and values of control limits;
step 7) respectively solving an approximate basis, a characteristic value and a characteristic vector for each type of sample;
step 8) calculating the SPE of each test sample respectively, wherein the SPE with the minimum difference value with the SPE of a certain type of sample is the prediction type of the test sample, the SPE is compared with the actual type, the accuracy is calculated, if the SPE meets the requirement, the step 9 is carried out, and if the SPE does not meet the requirement, the step 6 is carried out, and the parameter training is reset;
and 9) obtaining parameters, and finishing the algorithm.
Basic method of sparse kernel principal component analysis:
principal Component Analysis is a typical unsupervised algorithm, which is commonly used to solve the linear problem of the original space, and in order to solve the nonlinear problem of the original space by a linear method in the feature space, b.scholkopf et al propose Kernel Principal Component Analysis (KPCA).
Defining from an original space RnNonlinear mapping to feature space F:if a given sample X ═ X1,…,xN},xi∈RnThen pass throughThe mapping may obtain a set of vectorsAssuming that the set of vectors satisfiesThen the correlation matrix in the feature space is
If the set of vectorsThen can orderIt can be known thatSatisfies the condition of substituting inThe KPCA problem can be converted to solve a correlation matrix in feature spaceIs a feature vector
Wherein,is a linear combination of samples, such thatɑ＝[ɑ1,…,ɑN]TThen, thenWhen in useWhen the information can not be obtained explicitly, a kernel function is introduced and setFirst, it needs to calculate:
K＝ΦTΦ
where matrix K is a matrix of NxN, also known as a kernel matrix. The problem then translates into:
Kɑ＝Nλɑ
wherein, alpha is ═ alpha1,…,ɑN]. When in useThe process of centralization can be directly operated on K:
wherein,satisfy the requirement ofIs an NxN 1 matrix. Hypothesis obtainThe characteristic value lambda 1 is more than or equal to lambda 2 is more than or equal to … lambda n and the corresponding characteristic vector alpha thereof1,ɑ2,…,ɑN，Kth feature vector in feature spaceɑi,kTo representOf the kth feature vector of (1), byNormalized to obtainKth feature vector of variable x after normalizationThe projection of the direction is the kth principal component, and the formula is
The invention selects a common kernel function, namely a radial basis kernel function, to calculate. When selecting the radial basis kernel function, there is a significant over-learning problem, mainly because the feature space corresponding to the radial basis kernel function is wireless dimensional, and the number of principal components obtained by this method is independent of the dimension of a given sample, and is dependent on the number of samples. To solve this problem, a sparsification method, namely Sparse Kernel Principal Component Analysis (SKPCA), is introduced.
In kernel principal component analysis, feature vectors can be used as samplesIs shown asThat is to say feature vectorsIs a linear combination of all samples, which provides a thought for the sparseness of the feature vector. Solving by an approximate basis solving algorithmTo thereby obtain an approximation base ofThe thinning method of (1).
The concrete steps of the approximate base solving are as follows:
1. building a setTo approximate the maximum independent set, XA＝φ；
2. For k 2Solving a minimum value;
3. if the minimum value ≦ ε is obtained, the corresponding element is added to XAIn which otherwise is Xl。
Epsilon is a linear correlation truncation error, and the basis of an infinite dimensional space in a limited sample has almost no sparsity, so that an approximate calculation is selected.
4. And (5) returning to the step (2) until all calculations are completed, and ending the calculation.
Wherein, the solving process in the step 2 is as follows: order to
By Lagrange conditionsTo obtain-2K0+2K λ ═ 0, where K is0＝(k(x1,xk),…,k(xl,xk))TK is a square matrix, Kij＝k(xi,xj). When the kernel function is a Gaussian radial basis kernel function, the matrix K is positive definite to obtain lambdamin＝K-1K0This is the minimum point of f (λ).
Assuming a set of approximation bases to be foundBy using
The feature vector can be expressed asProblem is turned into
Ride on both sidesNote the bookTo obtain
It is known that the problem of the eigenvalues and eigenvectors, i.e.Order toThen
Derived by
Wherein K (m:) represents the m-th row of the kernel matrix K, K (: n) represents the n-th column, 1N＝[1,...,1]A row vector of 1 row and N columns is represented. Problem is also converted into (K)I)-1Ksα λ α is a typical eigenvalue and eigenvector problem.
Drawings
FIG. 1 is a flow chart of an embodiment of the method of the present invention;
FIG. 2 is the SPE difference between a portion of the test samples and each type;
FIG. 3 is a graph comparing the results of a certain portion of the test with the actual marked sample.
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.
As shown in fig. 1, a RdR scatter diagram identification method based on sparse kernel principal component analysis includes the following steps:
step 1) acquiring electrocardiosignals, carrying out filtering and denoising treatment on the electrocardiosignals, and extracting the position of an R wave peak value;
step 2) drawing RdR scatter diagram by MATLAB tool;
step 3) scaling RdR scatter diagram, converting into grey map, and normalizing image data to reduce calculated amount;
step 4), marking the obtained scatter diagram sample;
step 5) sampling samples, and randomly extracting 75% of data as training samples;
step 6) setting parameters, wherein the parameters comprise error parameters of the approximate basis, Gaussian kernel function parameters and values of control limits;
step 7) respectively solving an approximate basis, a characteristic value and a characteristic vector for each type of sample;
step 8) calculating the SPE of each test sample respectively, wherein the SPE with the minimum difference value with the SPE of a certain type of sample is the prediction type of the test sample, the SPE is compared with the actual type, the accuracy is calculated, if the SPE meets the requirement, the step 9 is carried out, and if the SPE does not meet the requirement, the step 6 is carried out, and the parameter training is reset;
and 9) obtaining parameters, and finishing the algorithm.
For convenience of description, the data of the database is schematically analyzed and described by using the electrocardiogram data in the MIT-BIH database. Fig. 2 is a graph (2) showing that the difference between a part of the test samples and each class of SPE calculation is the smallest, for example, a sample with the class "3" is selected, and the SPE with the class "3" is calculated as the smallest difference between the selected sample and each class of SPE calculation. FIG. 3 is a comparison graph of a part of test results of a certain time and actual marked samples according to the method of the present invention, wherein the parameter true is the actual sample category and the predicted is the category determined by the method of the present invention, and it can be seen that the accuracy is high.
Claims (6)
1. An RdR scatter diagram identification method based on sparse kernel principal component analysis is characterized by comprising the following steps:
step 1) acquiring electrocardiosignals, carrying out filtering and denoising treatment on the electrocardiosignals, and extracting the position of an R wave peak value;
step 2) drawing RdR scatter diagram by MATLAB tool;
step 3) scaling RdR scatter diagram, converting into grey map, and normalizing image data to reduce calculated amount;
step 4), marking the obtained scatter diagram sample;
step 5) sampling samples, and randomly extracting 75% of data as training samples;
step 6) setting parameters, wherein the parameters comprise error parameters of the approximate basis, Gaussian kernel function parameters and values of control limits;
step 7) respectively solving an approximate basis, a characteristic value and a characteristic vector for each type of sample;
step 8) calculating the SPE of each test sample respectively, wherein the SPE with the minimum difference value with the SPE of a certain type of sample is the prediction type of the test sample, the SPE is compared with the actual type, the accuracy is calculated, if the SPE meets the requirement, the step 9 is carried out, and if the SPE does not meet the requirement, the step 6 is carried out, and the parameter training is reset;
and 9) obtaining parameters, and finishing the algorithm.
2. The RdR scatter plot recognition method based on sparse kernel principal component analysis (SCMA), as claimed in claim 1, wherein: the kernel component analysis is an unsupervised machine learning algorithm.
3. The RdR scatter plot recognition method based on sparse kernel principal component analysis (SCMA), as claimed in claim 1, wherein: the RdR scatter diagram in the step 2) is a heart rate variability analysis method, which can show the trend of the heart rate variability analysis method along with the change of the heart rate variability analysis method.
4. The RdR scatter plot recognition method based on sparse kernel principal component analysis (SCMA), as claimed in claim 1, wherein: in the step 3), in order to reduce the calculation amount, the scatter diagram is converted into a gray scale diagram, and then the gray scale diagram is processed and identified.
5. The RdR scatter plot recognition method based on sparse kernel principal component analysis (SCMA), as claimed in claim 1, wherein: in the step 5), 75% of samples are randomly extracted for training and 25% of samples are randomly extracted for testing the performance of the classifier.
6. The RdR scatter plot recognition method based on sparse kernel principal component analysis (SCMA), as claimed in claim 1, wherein: in the step 7), in order to reduce redundant features and complexity of calculation, and an over-learning problem, the samples are thinned.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810884233.0A CN109119158A (en) | 2018-08-06 | 2018-08-06 | A kind of RdR scatter plot recognition methods based on sparse kernel principle component analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810884233.0A CN109119158A (en) | 2018-08-06 | 2018-08-06 | A kind of RdR scatter plot recognition methods based on sparse kernel principle component analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109119158A true CN109119158A (en) | 2019-01-01 |
Family
ID=64852639
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810884233.0A Pending CN109119158A (en) | 2018-08-06 | 2018-08-06 | A kind of RdR scatter plot recognition methods based on sparse kernel principle component analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109119158A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112562828A (en) * | 2020-12-10 | 2021-03-26 | 嘉兴芯识信息科技有限公司 | DICOM dynamic image and electrocardiographic waveform simultaneous representation method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106073755A (en) * | 2016-05-27 | 2016-11-09 | 成都信汇聚源科技有限公司 | The implementation method that in a kind of miniature holter devices, atrial fibrillation identifies automatically |
| CN108113670A (en) * | 2018-01-22 | 2018-06-05 | 江苏师范大学 | A kind of UAV system and control method of the control of multi-channel type induced brain wave |
-
2018
- 2018-08-06 CN CN201810884233.0A patent/CN109119158A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106073755A (en) * | 2016-05-27 | 2016-11-09 | 成都信汇聚源科技有限公司 | The implementation method that in a kind of miniature holter devices, atrial fibrillation identifies automatically |
| CN108113670A (en) * | 2018-01-22 | 2018-06-05 | 江苏师范大学 | A kind of UAV system and control method of the control of multi-channel type induced brain wave |
Non-Patent Citations (2)
| Title |
|---|
| 甘良志: ""核学习算法与集成方法研究"", 《中国博士学位论文全文数据库(信息科技辑)》 * |
| 陆宏伟 等: ""基于RdR新型散点图心率变异性研究"", 《生物医学工程学杂志》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112562828A (en) * | 2020-12-10 | 2021-03-26 | 嘉兴芯识信息科技有限公司 | DICOM dynamic image and electrocardiographic waveform simultaneous representation method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shi et al. | A hierarchical method based on weighted extreme gradient boosting in ECG heartbeat classification | |
| Fu et al. | Improvement motor imagery EEG classification based on regularized linear discriminant analysis | |
| Reasat et al. | Detection of inferior myocardial infarction using shallow convolutional neural networks | |
| Khalaf et al. | A novel technique for cardiac arrhythmia classification using spectral correlation and support vector machines | |
| CN109117730B (en) | Real-time electrocardiogram atrial fibrillation judgment method, device and system and storage medium | |
| Amrani et al. | Very deep feature extraction and fusion for arrhythmias detection | |
| Jung et al. | An arrhythmia classification method in utilizing the weighted KNN and the fitness rule | |
| CN106951914B (en) | Method for identifying vinegar variety by electronic nose for optimizing fuzzy identification vector extraction | |
| CN111631682B (en) | Physiological characteristic integration method and device based on trending analysis and computer equipment | |
| CN109157211A (en) | A kind of portable cardiac on-line intelligence monitoring diagnosis system design method | |
| CN112674782B (en) | Device and method for detecting epileptic-like electrical activity of epileptic during inter-seizure period | |
| Ozkok et al. | Convolutional neural network analysis of recurrence plots for high resolution melting classification | |
| Prakash et al. | A system for automatic cardiac arrhythmia recognition using electrocardiogram signal | |
| Li et al. | Automatic diagnosis of cardiac arrhythmia in electrocardiograms via multigranulation computing | |
| CN109316188A (en) | A kind of extracting method of migraine brain function connection mode | |
| CN107045624B (en) | Electroencephalogram signal preprocessing and classifying method based on maximum weighted cluster | |
| CN116109589A (en) | Image processing method applied to congestive heart failure diagnosis | |
| CN115089139A (en) | Personalized physiological parameter measuring method combining biological characteristic identification | |
| CN109119158A (en) | A kind of RdR scatter plot recognition methods based on sparse kernel principle component analysis | |
| Song et al. | Automatic identification of atrial fibrillation based on the modified Elman neural network with exponential moving average algorithm | |
| Bijinapalli | Prediction of Cardiac Arrhythmia using Random-Forest Machine Learning Algorithm | |
| CN108960113A (en) | A kind of heart rate variability recognition methods based on support vector machines | |
| Mohseni et al. | Heart arrhythmias classification via a sequential classifier using neural network, principal component analysis and heart rate variation | |
| CN113647954A (en) | Cardiovascular disease identification method, device and medium of two-channel hybrid network model | |
| Lee et al. | Comparative neural network based on template cluster for automated abnormal beat detection in electrocardiogram signals |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190101 |