CN109009222A - Intelligent evaluation diagnostic method and system towards heart disease type and severity - Google Patents
Intelligent evaluation diagnostic method and system towards heart disease type and severity Download PDFInfo
- Publication number
- CN109009222A CN109009222A CN201810630108.7A CN201810630108A CN109009222A CN 109009222 A CN109009222 A CN 109009222A CN 201810630108 A CN201810630108 A CN 201810630108A CN 109009222 A CN109009222 A CN 109009222A
- Authority
- CN
- China
- Prior art keywords
- heart disease
- severity
- data
- model
- disease type
- 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
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/08—Detecting organic movements or changes, e.g. tumours, cysts, swellings
- A61B8/0883—Detecting organic movements or changes, e.g. tumours, cysts, swellings for diagnosis of the heart
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/5215—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving processing of medical diagnostic data
- A61B8/5223—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving processing of medical diagnostic data for extracting a diagnostic or physiological parameter from medical diagnostic data
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/5284—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving retrospective matching to a physiological signal
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Medical Informatics (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Pathology (AREA)
- Radiology & Medical Imaging (AREA)
- Public Health (AREA)
- Heart & Thoracic Surgery (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Cardiology (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a kind of intelligent evaluation diagnostic method and system towards heart disease type and severity, this method include obtaining genius morbi data and Demographics data;Model-evaluation index, heart disease type and heart disease severity are analyzed and obtained using echocardiogram data reporting and patient Demographics data of the learning model to acquisition.The invention has the benefit that the correlation technique that maintenance data excavates carries out the operation such as data prediction, data screening to data, the selection of noise proportional when to Feature Selection, heart disease severity classification is carried out using Random Forest model to predict, and it compares and is analyzed with Naive Bayes Classifier, decision-tree model, the algorithm performance of BP neural network model and learning effect, it obtains effective research method, and proposes the prediction technique of the standard and cardiac surgery Operative risk that are classified to cardiac's coincident with severity degree of condition.
Description
Technical field
The present invention relates to the application of medical data and analysis technical fields, and in particular to a kind of towards heart disease type and tight
The intelligent evaluation diagnostic method and system of weight degree.
Background technique
Existing heart disease diagnosis method, clinician will report the heart completed to each patient according to heart disease diagnosis
The accurate judgement and heart disease Severity of popular name for type.This requires doctor that must have sturdy theoretical knowledge and many years
Clinical experience could complete.However, actually diagnosis in, often because the limitation of medical condition, doctor's energy, the time, experience,
The reasons such as diagnosis report description, cause the difficult diagnosis or deviation of doctor.
Summary of the invention
The purpose of the present invention is to provide a kind of intelligent evaluation diagnostic method towards heart disease type and severity and
System, it is existing because of reasons such as the limitation of medical condition, doctor's energy, time, experience, diagnosis report descriptions to solve, it causes
The problem of difficult diagnosis or deviation of doctor.
To achieve the above object, the technical scheme is that
On the one hand, a kind of intelligent evaluation diagnostic method towards heart disease type and severity, this method packet are provided
It includes:
Obtain genius morbi data and Demographics data;
Divided using echocardiogram data reporting and patient Demographics data of the learning model to acquisition
Analyse and obtain model-evaluation index, heart disease type and heart disease severity.
Further, the genius morbi data include LVEF, mitral leaflet thicken, bicuspid valve Echoenhance, active
Arteries and veins valve leaflet thickening, aorta petal leaflet, which are shown, owes that clear, aorta petal Echoenhance, opening of aortic valve be limited, aorta petal
There are gap, estimation aorta petal valve orifice area, left room, left room, right room, right ventricle, the dynamic degree diffusivities of locular wall slightly to lower when closure,
In systole phase right room visit and be distributed limitation tritubercular cycloid beam maximum backflow pressure difference, systole phase aorta petal forward blood flow accelerate
Maximum is backflowed pressure difference, aortic valve disease, aortic stenosis, aortic regurgitation, bicuspid valve hardening, mitral reflux, tricuspid
During valve backflows, Left ventricular systolic function slightly lowers, pulmonary hypertension, left ventricular hypertrophy, left ventricular enlargement, Myocardial damage can not rule out extremely
Few one kind.
Further, the Demographics data include at least one of name, gender and age.
Preferably, the learning model is Random Forest model, Naive Bayes Classifier, decision-tree model, BP nerve
One of network model.
Preferably, the learning model is Random Forest model.
Further, the model-evaluation index include Correctly, TP Rate, FP Rate, Precision,
At least one of Recall, F value, the area of Roc curve, Accuracy and threshold value.
Preferably, the heart disease type includes Myocardial damage (ischemic cardiomyopathy, hypertrophic cardiomyopathy, the expanding heart
Myopathy, other types cardiomyopathy), myocardial infarction, rheumatic heart disease, non-rheumatic valvular disease, atrial fibrillation, characteristic pulmonary hypertension,
Congenital heart disease, occupy-place (thrombus, pericardium occupy-place, other occupy-places), surgical postoperative (this kind of data we be all used as check number
At least one of according to come the accuracy rate that improves our machine learning).
Preferably, genius morbi data are obtained from echocardiogram report.
Preferably, the severity includes slight, moderate and severe.
On the other hand, provide a kind of intelligent evaluation diagnostic system towards heart disease type and severity, using appoint
The one above-mentioned intelligent evaluation diagnostic method towards heart disease type and severity.
The present invention has the advantage that
The related side excavated for the genius morbi data and Demographics data of cardiac, maintenance data
Method carries out the operation such as data prediction, data screening to data, and the selection of noise proportional, uses random forest when to Feature Selection
Model carried out heart disease severity classification prediction, and with Naive Bayes Classifier, decision-tree model, BP neural network mould
The algorithm performance and learning effect of type are compared and are analyzed, and obtain effective research method, and propose to heart disease
The standard of conditions of patients classification of severity and the prediction technique of cardiac surgery Operative risk;
The automatic diagnosis to cardiac is completed by the method for machine learning, used model has study energy
Power, can be with the forecasting accuracy of the increase lift scheme of training data.The cardiac risk assessment diagnostic model proposed, energy
The suggestion of diagnosis enough is provided for patient, medical assistance expert carries out medical diagnosis on disease, promotes the accuracy of diagnosis, helps medical institutions
Necessary medical resource is reserved, there is stronger practical application value.
Detailed description of the invention
Fig. 1 is echocardiogram report;
Fig. 2 is the line chart for changing influence of the number of iterations to accuracy;
Fig. 3 is to change the line chart of influence of the seed number to accuracy in the case where the number of iterations is 1000;
Fig. 4 is that the number of iterations is 1000, in the case that seed number is 1, changes the line chart of influence of the depth to accuracy rate;
Accuracy is 69.3878% when Fig. 5 is optimal under Random Forest model by model parameter tuning, iteration
Number 1000 times, remaining parameter keeps default value;
Fig. 6 is that S grades of ACC value is 75.817% (threshold value 0.2212) under the model of optimal random forest, and ROC is bent
Line area is 0.8765;
The ROC curve figure that Fig. 7 is S grades when random forest is optimal;
Fig. 8 is evaluation index contrast table.
Specific embodiment
The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention..
Embodiment 1
A kind of intelligent evaluation diagnostic method towards heart disease type and severity, comprising:
Obtain genius morbi data and Demographics data;
Divided using echocardiogram data reporting and patient Demographics data of the learning model to acquisition
Analyse and obtain model-evaluation index, heart disease type and heart disease severity.
It is of any of claims 1-9 towards heart disease type that the another aspect of the present embodiment discloses a kind of application
With the system of the intelligent evaluation diagnostic method of severity.
Embodiment 2
The genius morbi data include LVEF, mitral leaflet thickens, bicuspid valve Echoenhance, aorta petal leaflet increase
Thick, aorta petal leaflet show owe clear, aorta petal Echoenhance, when opening of aortic valve is limited, aorta petal is closed there are
Gap, estimation aorta petal valve orifice area, left room, left room, right room, right ventricle, the dynamic degree diffusivity of locular wall slightly lower, systole phase right room
Interior spy and distribution limitation tritubercular cycloid beam the maximum maximum that pressure difference, systole phase aorta petal forward blood flow accelerate of backflowing are backflowed pressure
Difference, aortic valve disease, aortic stenosis, aortic regurgitation, bicuspid valve hardening, mitral reflux, tritubercular cycloid, left room
Contractile function slightly lowers, pulmonary hypertension, left ventricular hypertrophy, left ventricular enlargement, Myocardial damage at least one of can not rule out.
The Demographics data include at least one of name, gender and age embodiment 3
The learning model is Random Forest model, Naive Bayes Classifier, decision-tree model, BP neural network mould
One of type.
The learning model is Random Forest model.
The model-evaluation index include Correctly, TP Rate, FP Rate, Precision, Recall, F value,
At least one of area, Accuracy and threshold value of Roc curve.
Embodiment 4
The heart disease type include Myocardial damage (ischemic cardiomyopathy, hypertrophic cardiomyopathy, dilated cardiomyopathy, its
His type cardiomyopathy), myocardial infarction, rheumatic heart disease, non-rheumatic valvular disease, atrial fibrillation, characteristic pulmonary hypertension, the congenital heart
Popular name for, occupy-place (thrombus, pericardium occupy-place, other occupy-places), surgical postoperative (this kind of data we all improved as verification data
At least one of the accuracy rate of our machine learning).
Genius morbi data are obtained from echocardiogram report.
The severity includes slight, moderate and severe.
Specific implementation process are as follows:
3.2 problem definition
The crucial wind being treated surgically between heart disease type and severity it was found that heart disease patient groups are admitted to hospital
Then dangerous factor establishes heart patient using improved algorithm of standing abreast at random and is admitted to hospital operation risk prediction model
Operation risk analysis of being admitted to hospital is carried out in order to the heart patient group accurately to covering gamut, set forth herein
Using integrated study, the model being independent of each other accordingly is established to specific area crowd, by whole in each submodel
Patient is for statistical analysis, and discovery heart disease patient groups, which are admitted to hospital, to be treated surgically between heart disease type and severity
Key risk factor, then establish heart patient using improved algorithm of standing abreast at random and be admitted to hospital operation risk prediction model;Most
Afterwards based on the model established, using clustering algorithm and principal component analytical method, identification is admitted to hospital there are high risk needs receives hand
The patient of art is described the medical features that it is reflected from Clinical symptoms level, reminds hospital, healthcare givers to mention in advance
It is preceding that medical resource is reserved to this kind of patient and is intervened in advance.
Specific experiment process: starting with from problem definition, then does data preparation, including collect data, typing, processing data
Process.By with doctor's communication exchange, feature construction is done according to the experience of doctor, then in data dirty data and noise into
Row processing.Data then are imported with weka software, introduce decision tree, random forest, neural network, model-naive Bayesian,
And model training is carried out, and the performance of model is improved by the methods of arameter optimization, finally find out that treatment accuracy is higher, speed
Faster model.Finally experimental result is assessed.
3.3 data preparation
1. data acquire
Under the guidance of doctor, under the precondition of testing result for not influencing patient, we are positive with patient
It links up and inquires its name, age and the echocardiogram for obtaining the patient after patient and doctor agree to, amount to and obtain 221 diseases
The echocardiogram report of trouble, and inquire that echocardiogram can be in the surgical effect of surgery and its which index Xiang doctor
The effectively reflection cardiopathic severity of sufferer.
2. data prediction
Whole features in the open in-heart operation under pulsating report for 221 sufferers being collected into have been carried out tentatively first
Statistics, all features amount to 520, weed out the feature unrelated with heart disease diagnosis and such as check number, instrument model, record
The attributes such as people, admission number, department, and to the report for thering is bulk information to lack, postoperative report information screens out.Following table is
The report of echocardiogram (" Demo07 " is classified as name in table, underground).
Under the guidance and help of doctor, our groups first, in accordance with heart disease type by all data classifications, and will diagnosis
Unknown data are rejected, and are specifically classified as follows:
Myocardial damage (ischemic cardiomyopathy, hypertrophic cardiomyopathy, dilated cardiomyopathy, other types cardiomyopathy)
Myocardial infarction
Rheumatic heart disease
Non- rheumatic valvular disease
Atrial fibrillation
Characteristic pulmonary hypertension
Congenital heart disease
Occupy-place (thrombus, pericardium occupy-place, other occupy-places)
Surgical postoperative (this kind of data we accuracys rate of our machine learning is all improved as verification data)
It is after doctor links up study the important attribute according to the experience of doctor to different type heart disease echocardiogram
It carries out building table.The dimensionality reduction for realizing first time, from 487 attributes (2 demographic attributes, 485 disease attributes, disease categories
Include in property:
1) M type and Two-dimensional Echocardiographic Features 130.
2) frequency spectrum and color Doppler feature 95.
3) feature 260 in conclusion are down to 173 attributes.
Full type feature that following table includes by the non-rheumatic valvular disease+ischemic cardiomyopathy of A class, full type number and
Its important feature.
Following table is to report that the part classifying typing that forms data carries out is united to ultrasound electrocardiogram according to different heart disease types
It counts table (" Demo07 " is classified as name in table, underground).
After first time dimensionality reduction, we renumber ill important attribute, will be under different heart disease types
Same alike result carry out Unified number and producing following table:
Existing attribute has 3 demographic attributes and 84, amounts to 87 attributes.Finally we obtain one tentatively
Experiment table (" Demo07 " is classified as name in table, underground), it is as follows:
We carry out the processing of missing values to the data information in table after obtaining the table, and eliminate the noise largely lacked
Feature, Characteristic Number is 79 at this time.Obtaining missing values treated experiment table, (" Demo07 " is classified as name in table, unjust
It opens).
Following table is the summary sheet to the processing method of each feature missing values:
3.4 latent structure
Latent structure mainly includes two aspect of feature selecting and Feature Dimension Reduction.
We have carried out preliminary system to whole features in the open in-heart operation under pulsating report (referring to Fig. 1) being collected into
Meter, all feature amount to 520, the first step we weed out the feature unrelated with heart disease diagnosis and such as check number, instrument type
Number, recorder, admission number, the attributes such as department, and remaining feature is divided into Demographics and genius morbi wherein people
Mouth statistics feature includes name, gender, age.
Under the guidance and help of doctor, our groups first, in accordance with heart disease type by all data classifications, and will diagnosis
Unknown data are rejected, and are specifically classified as follows:
Myocardial damage (ischemic cardiomyopathy, hypertrophic cardiomyopathy, dilated cardiomyopathy, other types cardiomyopathy)
Myocardial infarction
Rheumatic heart disease
Non- rheumatic valvular disease
Atrial fibrillation
Characteristic pulmonary hypertension
Congenital heart disease
Occupy-place (thrombus, pericardium occupy-place, other occupy-places)
Surgical postoperative (this kind of data we accuracys rate of our machine learning is all improved as verification data)
It is after doctor links up study the important attribute according to the experience of doctor to different type heart disease echocardiogram
It carries out building table.The dimensionality reduction for realizing first time, from 487 attributes (2 demographic attributes, 485 disease attributes, disease categories
Include in property:
1) M type and Two-dimensional Echocardiographic Features 130
2) frequency spectrum and color Doppler feature 95
3) feature 260 in conclusion are down to 173 attributes.
Full type feature, full type number and its important feature that following table includes by A class disease.
Following table (" Demo07 " is classified as name in table, underground) is according to different heart disease types to ultrasound electrocardiogram report
Accuse the part classifying typing that forms data carries out
After first time dimensionality reduction, we renumber ill important attribute, will be under different heart disease types
Same alike result carry out Unified number.The following figure is the number table carried out to each generic attribute:
Existing attribute has 3 demographic attributes and 84, amounts to 87 attributes.
Finally we obtain a preliminary experiment table (" Demo07 " is classified as name in table, underground), as follows:
3.5 model learning
3.5.1 tuning method
The heart disease coincident with severity degree of condition of sufferer is divided into five grades A, B, C by doctor according to ultrasonic cardiography diagram data,
D, E (severity successively weakens), wherein A, the sufferer suggestion of B grade cooperate surgical operation therapy.Meanwhile we are to improve
The accuracy of model analysis, and doctor is asked to assist for the heart disease severity of sufferer to be divided into three grades, respectively L is (light
Degree), M (moderate), S (severe) (severity successively enhances).Wherein the sufferer suggestion of S grade cooperates surgical operation therapy, together
When split data into ten folding cross validations and training test verification verifying two kinds, further to select the high model of accuracy.
Ten folding cross validations: it is used to testing algorithm accuracy, is common test method.Data set is divided into ten parts, choosing
Nine parts therein are used as training data, and portion is test data, carry out ten times, are tested by turns.Test obtains corresponding every time
Accuracy (or error rate).The accuracy (or error rate) of 10 achievements is averaged as the appraisal to arithmetic accuracy.
Training set, test set method we be will remaining data be half-and-half after postoperative corresponding pre-operative data removes in data
Point, then whole pre-operative datas is included into training set, preoperative and postoperative corresponding data should be checksum set, but WEKA is not supported, we
Verification data are incorporated to training set to improve the effect of model learning, hereinafter collectively referred to as training set.147 groups of all data are divided
For 81 groups of training set datas (including checksum set), 66 groups of test set data.
3.5.2 evaluation index
3-1 result index of table illustrates table
3.5.3 model training
Random Forest model
1) model selects: the random forest method under selection Weka, specifically: weka- > explorer- > Classify-
>Choose->trees->RandomForest.Model evaluation: as shown in Fig. 3-8.
Set five class of A, B, C, D, E and L, M, S three classes for hierarchy model, respectively using ten folding interior extrapolation methods and side set method into
Row test.
2) tuning model
It is classified as L, M by Experimental comparison's discovery, the accuracy of S three classes is better than being classified as A, B, C, five class of D, E, together
When ten folding cross validations so that the accuracy of model is improved, still select the method and three classes of ten folding cross validations
Classification.
3) arameter optimization
Number of iterations numiterations changes influence of the number of iterations to accuracy referring to fig. 2.
Seed number changes the line chart of influence of the seed number to accuracy in the case where the number of iterations is 1000, referring to
Fig. 3.
The depth of tree in the case that the number of iterations 1000, seed number are 1, changes influence of the depth to accuracy rate, referring to
Fig. 4.
Accuracy is 69.3878% when optimal under Random Forest model by model parameter tuning, the number of iterations
1000 times, remaining parameter keeps default value.Referring to Fig. 5
S grades of ACC value is 75.817% (threshold value 0.2212), ROC curve area under the model of optimal random forest
It is 0.8765, referring to Fig. 6.
S grades of ROC curve figure when random forest is optimal, referring to Fig. 7.
3.6 recruitment evaluation
3.6.1 evaluation index introduction
Following recruitment evaluation tables is formulated according to evaluation index
Evaluation index contrast table shown in Fig. 8 is produced according to evaluation index.
1. the accuracy of model-naive Bayesian is lower than its excess-three kind model, from TP in the analysis of different Model Diagnosis
Decision-tree model is above neural network model and Bayesian model and random gloomy from the point of view of rate, recall, accurary
Woods model.The correct classification number of four kinds of models is relatively high.
2. three kinds of disaggregated models have preferable accuracy rate as a whole, still in the diagnostic analysis based on prediction
It can be seen that from confusion matrix, cost may be biggish error rate.The accuracy rate of decision-tree model is better than neural network model, and
The accuracy rate of model-naive Bayesian is relatively low.Therefore, we select the higher decision-tree model of accuracy rate for examining automatically
It is disconnected.
3.6.2 experimental summary
In this part, we will assess the effect of machine learning, first we by the experience of doctor to feature into
Row building has obtained 79 feature vectors, then carries out model training using weka machine learning, is comparing examining for each model
Disconnected accuracy rate, F value etc., the decision-tree model for having chosen function admirable, being suitble to this project, and by the verifying for test set,
It was found that accuracy rate is up to 70.5873%, show that experiment effect is preferable.
It is studied by this, it has been found that:
1. the accuracy of random forest is higher in the case where data volume is sufficiently large, in the lesser situation of data volume, certainly
The accuracy of plan tree is higher.
2. influence it is cardiopathic it is several be mainly characterized by LVEF, aortic stenosis, Left ventricular systolic function lower etc..
3. the method for mask data has side to set method and ten folding interior extrapolation methods, ten foldings is selected to intersect in the case that data volume is less
Method, accuracy are higher.
Although above having used general explanation and specific embodiment, the present invention is described in detail, at this
On the basis of invention, it can be made some modifications or improvements, this will be apparent to those skilled in the art.Therefore,
These modifications or improvements without departing from theon the basis of the spirit of the present invention are fallen within the scope of the claimed invention.
Claims (10)
1. a kind of intelligent evaluation diagnostic method towards heart disease type and severity characterized by comprising
Obtain genius morbi data and Demographics data;
It is analyzed simultaneously using echocardiogram data reporting and patient Demographics data of the learning model to acquisition
Obtain model-evaluation index, heart disease type and heart disease severity.
2. the intelligent evaluation diagnostic method according to claim 1 towards heart disease type and severity, feature exist
In, genius morbi data include LVEF, mitral leaflet thicken, bicuspid valve Echoenhance, aorta petal leaflet thickening, master
Arterial valve leaflet is shown when owing limited clear, aorta petal Echoenhance, opening of aortic valve, aorta petal closure there are gap,
Estimation aorta petal valve orifice area, left room, left room, right room, right ventricle, the dynamic degree diffusivity of locular wall slightly lower, visit in systole phase right room
And distribution limitation tritubercular cycloid beam maximum backflow pressure difference, systole phase aorta petal forward blood flow accelerate maximum backflow pressure difference,
Aortic valve disease, aortic stenosis, aortic regurgitation, bicuspid valve hardening, mitral reflux, tritubercular cycloid, left room are received
Contracting function slightly lowers, pulmonary hypertension, left ventricular hypertrophy, left ventricular enlargement, Myocardial damage at least one of can not rule out.
3. the intelligent evaluation diagnostic method according to claim 1 towards heart disease type and severity, feature exist
In the Demographics data include at least one of name, gender and age.
4. the intelligent evaluation diagnostic method according to claim 1,2 or 3 towards heart disease type and severity, special
Sign is that the learning model is Random Forest model, Naive Bayes Classifier, decision-tree model, BP neural network model
One of.
5. the intelligent evaluation diagnostic method according to claim 4 towards heart disease type and severity, feature exist
In the learning model is Random Forest model.
6. the intelligent evaluation diagnostic method according to claim 5 towards heart disease type and severity, feature exist
In the model-evaluation index includes Correctly, TP Rate, FP Rate, Precision, Recall, F value, Roc curve
At least one of area, Accuracy and threshold value.
7. the intelligent evaluation diagnostic method according to claim 5 towards heart disease type and severity, feature exist
In the heart disease type includes Myocardial damage (ischemic cardiomyopathy, hypertrophic cardiomyopathy, dilated cardiomyopathy, other types
Cardiomyopathy), myocardial infarction, rheumatic heart disease, non-rheumatic valvular disease, atrial fibrillation, characteristic pulmonary hypertension, congenital heart disease,
Occupy-place (thrombus, pericardium occupy-place, other occupy-places), surgical postoperative (this kind of data we all improve us as verification data
At least one of the accuracy rate of machine learning).
8. the intelligent evaluation diagnostic method according to claim 5 towards heart disease type and severity, feature exist
In the acquisition genius morbi data from echocardiogram report.
9. the intelligent evaluation diagnostic method according to claim 5 towards heart disease type and severity, feature exist
In the severity includes slight, moderate and severe.
10. a kind of intelligent evaluation diagnostic system towards heart disease type and severity, which is characterized in that apply claim
Intelligent evaluation diagnostic method described in any one of 1-9 towards heart disease type and severity.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810630108.7A CN109009222A (en) | 2018-06-19 | 2018-06-19 | Intelligent evaluation diagnostic method and system towards heart disease type and severity |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810630108.7A CN109009222A (en) | 2018-06-19 | 2018-06-19 | Intelligent evaluation diagnostic method and system towards heart disease type and severity |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109009222A true CN109009222A (en) | 2018-12-18 |
Family
ID=64609592
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810630108.7A Pending CN109009222A (en) | 2018-06-19 | 2018-06-19 | Intelligent evaluation diagnostic method and system towards heart disease type and severity |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109009222A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110200620A (en) * | 2019-06-06 | 2019-09-06 | 中山大学孙逸仙纪念医院 | Multi-modal cardiac data detection method, device, storage medium and terminal device |
CN110289095A (en) * | 2019-06-28 | 2019-09-27 | 青岛百洋智能科技股份有限公司 | A kind of fracture of neck of femur clinic intelligence aided decision method and system |
CN110706812A (en) * | 2019-09-29 | 2020-01-17 | 医渡云(北京)技术有限公司 | Medical index time sequence prediction method, device, medium and electronic equipment |
CN111180064A (en) * | 2019-12-25 | 2020-05-19 | 北京亚信数据有限公司 | Evaluation method and device for auxiliary diagnosis model and computing equipment |
CN111191773A (en) * | 2020-01-07 | 2020-05-22 | 泰康保险集团股份有限公司 | Method, apparatus, medium, and device for updating network parameters of information prediction network |
CN111419283A (en) * | 2020-04-08 | 2020-07-17 | 上海长征医院 | Ultrasonic analysis system for evaluating severity of new coronary pneumonia patient |
CN111507964A (en) * | 2020-04-17 | 2020-08-07 | 中山仰视科技有限公司 | Integrated learning grading method for new coronary pneumonia, electronic equipment and storage medium |
CN111930948A (en) * | 2020-09-08 | 2020-11-13 | 平安国际智慧城市科技股份有限公司 | Information collection and classification method and device, computer equipment and storage medium |
WO2021249048A1 (en) * | 2020-06-10 | 2021-12-16 | 北京核信锐视安全技术有限公司 | Coronavirus disease 2019 (covid-19) risk prompt apparatus, method and system based on ultrasonic imaging |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1194045A (en) * | 1995-07-25 | 1998-09-23 | 好乐思治疗公司 | Computer assisted methods for diagnosing diseases |
CN1748217A (en) * | 2002-12-18 | 2006-03-15 | Ge医疗系统环球技术有限公司 | Enhanced computer-assisted medical data processing system and method |
CN1914617A (en) * | 2004-02-03 | 2007-02-14 | 美国西门子医疗解决公司 | Systems and methods for automated diagnosis and decision support for heart related diseases and conditions |
CN104736055A (en) * | 2012-05-30 | 2015-06-24 | 瑞思迈传感器技术有限公司 | Method and apparatus for monitoring cardio-pulmonary health |
US20150376704A1 (en) * | 2009-12-09 | 2015-12-31 | Cleveland Heartlab, Inc. | Biomarker assay for diagnosis and classification of cardiovascular disease |
CN105228508A (en) * | 2013-03-08 | 2016-01-06 | 新加坡健康服务有限公司 | A kind of system and method measured for the risk score of classifying |
US20170127964A1 (en) * | 2014-06-17 | 2017-05-11 | University Of Virginia Patent Foundation | System and method for improved cardiac rhythm classification from the time between heart beat intervals using non-linear dynamics |
CN106874663A (en) * | 2017-01-26 | 2017-06-20 | 中电科软件信息服务有限公司 | Cardiovascular and cerebrovascular disease Risk Forecast Method and system |
CN107145756A (en) * | 2017-05-17 | 2017-09-08 | 上海辉明软件有限公司 | A kind of stroke types Forecasting Methodology and device |
US20180025133A1 (en) * | 2016-07-21 | 2018-01-25 | MedAvante-ProPhase, Inc. | System for guiding clinical decisions using improved processing of data collected during a clinical trial |
-
2018
- 2018-06-19 CN CN201810630108.7A patent/CN109009222A/en active Pending
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1194045A (en) * | 1995-07-25 | 1998-09-23 | 好乐思治疗公司 | Computer assisted methods for diagnosing diseases |
CN1748217A (en) * | 2002-12-18 | 2006-03-15 | Ge医疗系统环球技术有限公司 | Enhanced computer-assisted medical data processing system and method |
CN1914617A (en) * | 2004-02-03 | 2007-02-14 | 美国西门子医疗解决公司 | Systems and methods for automated diagnosis and decision support for heart related diseases and conditions |
US20150376704A1 (en) * | 2009-12-09 | 2015-12-31 | Cleveland Heartlab, Inc. | Biomarker assay for diagnosis and classification of cardiovascular disease |
CN104736055A (en) * | 2012-05-30 | 2015-06-24 | 瑞思迈传感器技术有限公司 | Method and apparatus for monitoring cardio-pulmonary health |
CN105228508A (en) * | 2013-03-08 | 2016-01-06 | 新加坡健康服务有限公司 | A kind of system and method measured for the risk score of classifying |
US20170127964A1 (en) * | 2014-06-17 | 2017-05-11 | University Of Virginia Patent Foundation | System and method for improved cardiac rhythm classification from the time between heart beat intervals using non-linear dynamics |
US20180025133A1 (en) * | 2016-07-21 | 2018-01-25 | MedAvante-ProPhase, Inc. | System for guiding clinical decisions using improved processing of data collected during a clinical trial |
CN106874663A (en) * | 2017-01-26 | 2017-06-20 | 中电科软件信息服务有限公司 | Cardiovascular and cerebrovascular disease Risk Forecast Method and system |
CN107145756A (en) * | 2017-05-17 | 2017-09-08 | 上海辉明软件有限公司 | A kind of stroke types Forecasting Methodology and device |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110200620A (en) * | 2019-06-06 | 2019-09-06 | 中山大学孙逸仙纪念医院 | Multi-modal cardiac data detection method, device, storage medium and terminal device |
CN110289095A (en) * | 2019-06-28 | 2019-09-27 | 青岛百洋智能科技股份有限公司 | A kind of fracture of neck of femur clinic intelligence aided decision method and system |
CN110706812A (en) * | 2019-09-29 | 2020-01-17 | 医渡云(北京)技术有限公司 | Medical index time sequence prediction method, device, medium and electronic equipment |
CN111180064A (en) * | 2019-12-25 | 2020-05-19 | 北京亚信数据有限公司 | Evaluation method and device for auxiliary diagnosis model and computing equipment |
CN111191773A (en) * | 2020-01-07 | 2020-05-22 | 泰康保险集团股份有限公司 | Method, apparatus, medium, and device for updating network parameters of information prediction network |
CN111419283A (en) * | 2020-04-08 | 2020-07-17 | 上海长征医院 | Ultrasonic analysis system for evaluating severity of new coronary pneumonia patient |
CN111419283B (en) * | 2020-04-08 | 2023-11-14 | 上海长征医院 | Ultrasonic analysis system for evaluating disease severity of patient with new coronaries |
CN111507964A (en) * | 2020-04-17 | 2020-08-07 | 中山仰视科技有限公司 | Integrated learning grading method for new coronary pneumonia, electronic equipment and storage medium |
WO2021249048A1 (en) * | 2020-06-10 | 2021-12-16 | 北京核信锐视安全技术有限公司 | Coronavirus disease 2019 (covid-19) risk prompt apparatus, method and system based on ultrasonic imaging |
CN111930948A (en) * | 2020-09-08 | 2020-11-13 | 平安国际智慧城市科技股份有限公司 | Information collection and classification method and device, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109009222A (en) | Intelligent evaluation diagnostic method and system towards heart disease type and severity | |
Kusunose et al. | A deep learning approach for assessment of regional wall motion abnormality from echocardiographic images | |
Gandhi et al. | Automation, machine learning, and artificial intelligence in echocardiography: a brave new world | |
Clifford et al. | Classification of normal/abnormal heart sound recordings: The PhysioNet/Computing in Cardiology Challenge 2016 | |
US8295576B2 (en) | System and method for automated medical diagnostic interpretation and report generation | |
Avci et al. | An intelligent diagnosis system based on principle component analysis and ANFIS for the heart valve diseases | |
US20080103403A1 (en) | Method and System for Diagnosis of Cardiac Diseases Utilizing Neural Networks | |
KR102439082B1 (en) | Method for predicting liver disease of ordinary perseon using ecg analysis data based on deep running | |
Pal et al. | Data mining approach for coronary artery disease screening | |
Levine et al. | The risk of having additional obstructive lesions in neonatal coarctation of the aorta | |
Kovalchin et al. | Echocardiographic hemodynamic and morphometric predictors of survival after two-ventricle repair in infants with critical aortic stenosis | |
Khaled et al. | Employing time-domain methods and poincaré plot of heart rate variability signals to detect congestive heart failure | |
Gautam et al. | Contemporary applications of machine learning for device therapy in heart failure | |
Chen et al. | Heart murmurs clustering using machine learning | |
Zhang et al. | Automatic assessment of mitral regurgitation severity using the mask R-CNN algorithm with color doppler echocardiography images | |
Al-Gburi et al. | Optical disk segmentation in human retina images with golden eagle optimizer | |
Podgorelec et al. | Self-adapting evolutionary decision support model | |
Podforelec et al. | Evolutionary construction of medical decision trees | |
Cornely et al. | Heart murmur detection using wavelet time scattering and support vector machines | |
Stasis et al. | A multiple decision trees architecture for medical diagnosis: The differentiation of opening snap, second heart sound split and third heart sound | |
CN117672460A (en) | Mitral valve repair difficulty assessment method and system based on ultrasonic and machine learning | |
Rajeshwari et al. | Detection of phonocardiogram event patterns in mitral valve prolapse: An automated clinically relevant explainable diagnostic framework | |
Podgorelec et al. | Medical diagnosis prediction using genetic programming | |
Bohacik et al. | Fuzzy rule-based system applied to risk estimation of cardiovascular patients | |
lee Colunga | Mathematical Analysis of Cardiovascular Hemodynamics in Disease |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181218 |