CN110176304A - A kind of method and device of determining fracture in patients with diabetes risk - Google Patents
A kind of method and device of determining fracture in patients with diabetes risk Download PDFInfo
- Publication number
- CN110176304A CN110176304A CN201910317945.9A CN201910317945A CN110176304A CN 110176304 A CN110176304 A CN 110176304A CN 201910317945 A CN201910317945 A CN 201910317945A CN 110176304 A CN110176304 A CN 110176304A
- Authority
- CN
- China
- Prior art keywords
- fracture
- diabetic
- sampled data
- model
- risk
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- 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
- 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
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Theoretical Computer Science (AREA)
- Primary Health Care (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The embodiment of the invention discloses a kind of method and devices of determining fracture in patients with diabetes risk, wherein method includes: to obtain the sampled data of target diabetic, and the sampled data of target diabetic is inputted into object module, obtain the value-at-risk of target fracture in patients with diabetes;Wherein, sampled data includes at least one in blood sample achievement data and clinical indices data, and object module is obtained according to the sampled data and fracture medical treatment result training of multiple first diabetics.In the embodiment of the present invention, the risk that target diabetic suffers from fracture is analyzed by blood sample index and/or clinical indices, it may be implemented to predict the risk of bone fracture of diabetic, and the prediction effect predicted the risk of bone fracture of diabetic can be made preferable.
Description
Technical field
The present invention relates to medical data processing technology field more particularly to a kind of sides of determining fracture in patients with diabetes risk
Method and device.
Background technique
Diabetic osteoporosis refers to diabetic because of bone amount reduction, bone in the unit volume caused by suffering from diabetes
Tissue microstructure changes, bone strength reduces, the increased symptom of bone brittleness, be diabetes skeletal system important complication it
One.However, diabetic osteoporosis is usually not significantly characterized in the early stage of patient, when illness presentation occurs
When, patient may have occurred serious sclerotin pain, bone deformity, or even fracture, thus to diabetic's
Quality of life brings serious influence.Therefore, the risk of bone fracture for predicting diabetic dredges prevention diabetic keratopathy sclerotin
Pine, the quality of life for improving diabetic are very important.
Risk of bone fracture factor appraisal procedure (Fracture Risk Assessment Tool, FRAX) is current using most
Extensive risk of bone fracture appraisal procedure can carry out the risk of bone fracture for being predicted object coming 10 years using this kind of appraisal procedure
Prediction.Specifically, FRAX method can be based on being predicted the bone density achievement data of object to the fracture for being predicted object
Risk is predicted, if the bone density index for being predicted object is higher, the risk for being predicted object trouble fracture is smaller, if by
It predicts that the bone density index of object is lower, then it is larger to be predicted the risk that object suffers from fracture.However, applicant passes through multinomial reality
Issue after examination and approval existing, FRAX method is to the prediction effect of fracture in patients with diabetes risk and bad, for example, the bone of type 2 diabetic patient is strong
It is usually lower to spend achievement data, and bone density achievement data is usually normal or relatively high, therefore using FRAX method to 2 types
The risk of bone fracture of diabetic predicts that the risk of bone fracture that will lead to diabetes B is underestimated;Therefore, it is directed to glycosuria
For patient, FRAX method can not predict risk of bone fracture well.
To sum up, a kind of method for needing determining fracture in patients with diabetes risk at present, to realize to diabetic's
Risk of bone fracture is predicted.
Summary of the invention
The embodiment of the present invention provides a kind of method and device of determining fracture in patients with diabetes risk, to realize to glycosuria
The risk of bone fracture of patient is predicted.
In a first aspect, a kind of method of determining fracture in patients with diabetes risk provided in an embodiment of the present invention, the method
Include:
The sampled data of target diabetic is obtained, the sampled data of the target diabetic includes the target
At least one of in the blood sample achievement data and clinical indices data of diabetic;By the sampling of the target diabetic
Data input object module, obtain the value-at-risk of the target fracture in patients with diabetes;The object module is according to multiple
What the sampled data of one diabetic and fracture medical treatment result training obtained.
In above-mentioned technical proposal, object module is obtained by using blood sample achievement data and/or the training of clinical indices data,
It may be implemented to predict the risk of bone fracture of diabetic, and can to carry out the risk of bone fracture of diabetic pre-
The prediction effect of survey is preferable.
Optionally, the object module is the sampled data and fracture medical treatment result instruction according to multiple first diabetics
It gets, comprising: obtained at least using the sampled data and fracture medical treatment result training of the multiple first diabetic
One prediction model;Obtain the sampled data and fracture medical treatment result of multiple second diabetics;It is directed to described at least one
Each prediction model in a prediction model executes: the sampled data input of the multiple second diabetic is described pre-
It surveys model and obtains the value-at-risk of the multiple second fracture in patients with diabetes;According to the fracture of the multiple second diabetic
The value-at-risk of medical treatment result and the multiple second fracture in patients with diabetes, determines the prediction effect of the prediction model;By institute
Prediction effect is best at least one prediction model prediction model is stated as object module.
In above-mentioned technical proposal, by that will train to obtain the prediction model conduct that prediction effect is best in multiple preset models
Object module can make the prediction effect of object module preferable, so that the target sugar predicted based on object module
The risk of bone fracture for urinating patient is more accurate.
Optionally, at least one described prediction model includes Logic Regression Models, supporting vector machine model, random forest mould
At least one of in type and neural network model;Use the sampled data and fracture diagnosis and treatment knot of the multiple first diabetic
Fruit training obtains at least one prediction model, includes at least one of the following: the sampling using the multiple first diabetic
Data and fracture medical treatment result pass through the training of logistic regression algorithm and obtain the Logic Regression Models;Use the multiple first sugar
The sampled data for urinating patient obtains the supporting vector machine model by algorithm of support vector machine training with fracture medical treatment result;
Multiple decision-tree models are obtained using the sampled data and fracture medical treatment result of the multiple first diabetic, according to described
Multiple decision-tree models obtain the Random Forest model;Use the sampled data and fracture of the multiple first diabetic
Medical treatment result obtains the neural network model by the training of neural network machine algorithm.
In above-mentioned technical proposal, Logic Regression Models, supporting vector machine model, Random Forest model and neural network model
It is to be obtained by machine learning method training, on the one hand, predict diabetic's by using the method for machine learning
Risk of bone fracture can be realized with relatively advanced technology and be predicted the risk of bone fracture of patient of diabetes, to promote at data
Manage the development in medical domain;On the other hand, at least one fracture in patients with diabetes is established by using the method for machine learning
The personalization to vary with each individual precisely diagnosis may be implemented, for example, can be based on target diabetic's in the prediction model of risk
Sampled data determines object module, so that more accurate to the prediction result of target diabetic.
Optionally, the blood sample index includes fasting blood-glucose, postprandial 1 hour blood glucose, postprandial 2 hours blood glucose, HbAles
Albumen;The clinical indices include gender, at the age, occupation, height, weight, smoke, drink, family history, the course of disease, body temperature, arteries and veins
It fights, breathe, any one in blood pressure or any multinomial.
Optionally, the blood sample index further includes following any one or any multinomial: total cholesterol, triglycerides, load
Lipoprotein A, apolipoprotein B, high-density lipoprotein cholesterol, low density lipoprotein cholesterol.
In above-mentioned technical proposal, by can be adapted for using a large amount of clinical indices and physiochemical indice as sampled data
The special circumstances of diabetic, so as to solve existing FRAX method as be included in index it is less caused by unpredictable sugar
The technical issues of urinating patient's risk of bone fracture preferably carries out the prediction of fracture in patients with diabetes risk.
It optionally, further include bone density achievement data and medical history achievement data in the sampled data;The bone density refers to
Mark includes bone density, electronic constant computer tomography CT bone density;The medical history index includes the past History of Bone Fracture, parent
Hip Fracture history.
In above-mentioned technical proposal, by by bone density achievement data, medical history achievement data, physiochemical indice and clinical indices into
Row combines, and sampled data can be made more comprehensive, further improve the accuracy of prediction model.
Second aspect, a kind of device of determining fracture in patients with diabetes risk provided in an embodiment of the present invention, described device
Include:
Module is obtained, for obtaining the sampled data of target diabetic, the hits of the target diabetic
According at least one in blood sample achievement data and clinical indices data including the target diabetic;
Prediction module obtains the target for the sampled data of the target diabetic to be inputted object module
The value-at-risk of fracture in patients with diabetes;The object module is examined according to the sampled data and fracture of multiple first diabetics
Treat what result training obtained.
Optionally, described device further includes training module, and the training module is used for: using the multiple first diabetes
The sampled data of patient and fracture medical treatment result training obtain at least one prediction model;It is directed at least one described prediction mould
Each prediction model in type executes: the sampled data of the multiple second diabetic being inputted the prediction model and is obtained
To the value-at-risk of the multiple second fracture in patients with diabetes;According to the fracture medical treatment result of the multiple second diabetic
With the value-at-risk of the multiple second fracture in patients with diabetes, the prediction effect of the prediction model is determined;By described at least one
The best prediction model of prediction effect is as object module in a prediction model.
Optionally, at least one described prediction model includes Logic Regression Models, supporting vector machine model, random forest mould
Type and neural network model;The training module is at least one of following for executing: using the multiple first diabetic
Sampled data and fracture medical treatment result by logistic regression algorithm training obtain the Logic Regression Models;Using the multiple
The sampled data and fracture medical treatment result of first diabetic obtains the supporting vector by algorithm of support vector machine training
Machine model;Multiple decision-tree models are obtained using the sampled data and fracture medical treatment result of the multiple first diabetic,
The Random Forest model is obtained according to the multiple decision-tree model;Use the hits of the multiple first diabetic
The neural network model is obtained by the training of neural network machine algorithm according to fracture medical treatment result.
The third aspect, the embodiment of the invention also provides a kind of computer readable storage mediums, including instruction, when it is being counted
When being run on calculation machine so that computer execute as above-mentioned first aspect or first aspect arbitrarily as described in determination diabetic's bone
Roll over the method for risk.
Fourth aspect, the embodiment of the invention also provides a kind of computer program products, when run on a computer,
So that computer execute as first aspect or first aspect arbitrarily as described in determination fracture in patients with diabetes risk method.
These aspects or other aspects of the application can more straightforward in the following description.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly introduced, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill in field, without any creative labor, it can also be obtained according to these attached drawings
His attached drawing.
Fig. 1 is a kind of corresponding process signal of method of determining fracture in patients with diabetes risk provided in an embodiment of the present invention
Figure;
Fig. 2 is a kind of corresponding flow diagram of method of determining object module provided in an embodiment of the present invention;
Fig. 3 is a kind of structural schematic diagram of decision tree provided in an embodiment of the present invention;
Fig. 4 is a kind of structural schematic diagram of the device of determining fracture in patients with diabetes risk provided in an embodiment of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
It is described in detail to one step, it is clear that described embodiments are only a part of the embodiments of the present invention, rather than whole implementation
Example.Based on the embodiments of the present invention, obtained by those of ordinary skill in the art without making creative efforts
All other embodiment, shall fall within the protection scope of the present invention.
Fig. 1 is a kind of corresponding process signal of method of determining fracture in patients with diabetes risk provided in an embodiment of the present invention
Figure, this method comprises:
Step 101, the sampled data of target diabetic is obtained.
In the embodiment of the present invention, the acquisition modes of the sampled data of target diabetic can there are many, one kind can
In the implementation of energy, the sampled data of target diabetic can obtain target glycosuria in the first preset time period to pass through
Patient obtains in the diagnosis records of multiple hospitals, wherein the first preset time period can by those skilled in the art according to
Experience is configured, and is specifically not construed as limiting.By taking the first preset time period is 3 months as an example, if target diabetic is nearest
3 years in see a doctor respectively in first hospital, second hospital, then can obtain target diabetic respectively the first of first hospital
Diagnosis records and the second diagnosis records in second hospital, and can be obtained by the first diagnosis records of analysis and the second diagnosis records
To the sampled data of target diabetic.
In specific implementation, the sampled data of target diabetic may include the blood sample index number of target diabetic
According to in clinical indices data at least one of, wherein blood sample index can refer to the indices in collected blood, face
Clinical examination index when bed index can refer to medical.In one example, blood sample index may include fasting blood-glucose, it is postprandial
1 hour blood glucose, postprandial 2 hours blood glucose and glycosylated hemoglobin, aforementioned four index are the finger directly related with diabetic disorders
Mark can be between qualitative characterization's diabetes and fracture by the risk based on aforementioned four index analysis fracture in patients with diabetes
Incidence relation.By taking fasting blood-glucose as an example, in general, the normal range (NR) of fasting blood sugar can be 3.5-6.2mmol/L, if
It determines that the fasting blood sugar of patient is greater than or equal to 7.0mmol/L by detecting blood sample, then can determine that patient suffers from diabetes;Phase
Ying Di, if diabetic A has high risk of bone fracture, diabetic B has low risk of bone fracture, then is suffered from by analysis of diabetes
The difference of the fasting blood sugar of person A and diabetic B can determine being associated between risk of bone fracture and diabetes.It needs
Bright, above-mentioned example only describes the process based on a kind of blood sample index (i.e. fasting blood-glucose) analysis risk of bone fracture, the present invention
In embodiment, by the aforementioned four blood sample index of comprehensive analysis, the relationship between risk of bone fracture and diabetic disorders can be determined,
So as to the risk of Accurate Prediction fracture in patients with diabetes.
Optionally, blood sample index can also include total cholesterol, triglycerides, aPoA, apolipoprotein B, high density
Any one or any multinomial in lipoprotein cholesterol and low density lipoprotein cholesterol.For example, in example one, blood
Sample index may include fasting blood-glucose, postprandial 1 hour blood glucose, postprandial 2 hours blood glucose, glycosylated hemoglobin and total cholesterol;?
In example two, blood sample index may include fasting blood-glucose, postprandial 1 hour blood glucose, postprandial 2 hours blood glucose, glycosylated hemoglobin, sweet
Oily three esters and low density lipoprotein cholesterol;In example three, blood sample index may include fasting blood-glucose, postprandial 1 hour blood glucose,
Postprandial 2 hours blood glucose, glycosylated hemoglobin, total cholesterol, triglycerides, aPoA, apolipoprotein B, high-density lipoprotein
Cholesterol and low density lipoprotein cholesterol.
Specifically, total cholesterol index, triglycerides index, aPoA index, apolipoprotein B index, high density
Lipoprotein cholesterol index and low density lipoprotein cholesterol index are indexs relevant to fat and cardiovascular and cerebrovascular disease, are passed through
It, can be with qualitative characterization's diabetes based on the risk of one or more index analysis fracture in patients with diabetes in above-mentioned six indexs
With the incidence relation between fat and cardiovascular and cerebrovascular disease.By taking total cholesterol as an example, in general, the normal range (NR) of total cholesterol
It can be 2.85~5.69mmol/L, if determining that the total cholesterol of patient is greater than or equal to 7.4mmol/L by detecting blood sample,
It is fat can to determine that patient suffers from, easily causes cardiovascular and cerebrovascular disease;Correspondingly, if diabetic C has high risk of bone fracture, glycosuria
Patient D has low risk of bone fracture, then, can be with by the difference of the total cholesterol of analysis of diabetes patient A and diabetic B
Determine being associated between risk of bone fracture and obesity and cardiovascular and cerebrovascular disease.It is based on it should be noted that above-mentioned example only describes
Total cholesterol analyze risk of bone fracture process, if blood sample index include it is multiple in aforementioned four index and six indexs, can
To determine between risk of bone fracture and diabetic disorders, obesity and cardiovascular and cerebrovascular disease by the multiple blood sample indexs of comprehensive analysis
Relationship, so as to the risk of Accurate Prediction fracture in patients with diabetes.
Optionally, in order to improve the accuracy for predicting fracture in patients with diabetes risk, blood sample index may include that total gallbladder is red
Element, albumin, globulin, blood cholinesterase, a-L fucosidase, creatine kinase, creatine kinase isozyme, lactic dehydrogenase,
A- hydroxybutyric dehydrogenase, transferrins, cystatin C, total cholesterol, triglyceride, aPoA, carries rouge egg at blood amylase
White B, high-density lipoprotein cholesterol, low density lipoprotein cholesterol, glycosylated hemoglobin, c reactive protein, urine Microalbunin
White, fasting blood-glucose, postprandial 1 hour blood glucose, postprandial 2 hours blood glucose etc..It, can by using comprehensive blood sample index as training data
To improve the accuracy of prediction model, to improve the accuracy of the risk of prediction fracture in patients with diabetes.
Further, clinical indices may include gender, at the age, height, weight, smoke, drink, is in family history any
One or any multinomial.Optionally, in order to improve the comprehensive of clinical indices, clinical indices can also include height, the course of disease, body
Temperature, pulse, breathing, blood pressure, potassium, sodium, chlorine, total CO 2, anionic gap, calcium, urea/creatinine, glucose, osmotic pressure, β
Hydroxybutyric acid, uric acid, alkaline phosphatase, glutamic-oxalacetic transaminease, glutamic-pyruvic transaminase, any one in glutamyl transpeptidase or any
It is multinomial.In this way, can substantially more understand target diabetic's by using a large amount of clinical indices as sampled data
Physical condition, so that more fully to the analysis of target diabetic, the bone of target diabetic is better anticipated
Roll over risk.
In the embodiment of the present invention, sampled data further includes bone density achievement data and/or medical history achievement data, wherein bone
Density index may include bone density, electronic constant computed tomography (Computed Tomography, CT) bone density
Deng medical history index may include the past History of Bone Fracture, parent's Hip Fracture history.Correspondingly, the sampled data of target diabetic
May include the bone density data of target diabetic, quantitive CT bone density data, target diabetic history bone
Roll over number, history fracture the time, history fracture factor, target diabetic parent hip history fracture number, history
Fracture time, history fracture factor etc..
In one example, sampled data may include blood sample achievement data, clinical indices data, bone density achievement data
With medical history achievement data, i.e., predict diabetic risk of bone fracture when, by bone density achievement data, medical history achievement data,
Physiochemical indice and clinical indices are combined;Due to considering the index of multiple dimensions, so that sampled data is more
Comprehensively, it can be realized the Accurate Prediction to target fracture in patients with diabetes risk.
Step 102, the sampled data of target diabetic is inputted into object module, obtains target fracture in patients with diabetes
Value-at-risk.
In one example, the value-at-risk for the target fracture in patients with diabetes that object module is predicted can for " 0 " and
Any one in " 1 ", if the value-at-risk of target fracture in patients with diabetes is " 0 ", it can be said that improving eyesight mark diabetic is
Fracture low-risk diabetic, if the value-at-risk of target fracture in patients with diabetes is " 1 ", it can be said that improving eyesight mark patient of diabetes
Person is high fracture risk diabetic.
It should be noted that above-mentioned example is only a kind of illustrative simple declaration, cited by target patient of diabetes
The value-at-risk of person's fracture is merely for convenience and purposes of illustration of scheme, does not constitute the restriction to scheme, in specific implementation, target glycosuria
The value-at-risk of patient's fracture other can also be worth, for example can be the arbitrary value in [0,1], specifically be not construed as limiting.With target
For the value-at-risk of fracture in patients with diabetes is the arbitrary value in [0,1], if the value-at-risk of target fracture in patients with diabetes is higher,
It can be said that a possibility that improving eyesight mark fracture in patients with diabetes risk, is bigger.
In a step 102, object module can be the sampled data and fracture diagnosis and treatment according to multiple first diabetics
As a result training obtains, and the specific implementation of determining object module is described below.
Fig. 2 is a kind of corresponding flow diagram of method of determining object module provided in an embodiment of the present invention, this method
Include:
Step 201, the sampled data and fracture medical treatment result of multiple first diabetics are obtained.
In the embodiment of the present invention, the sampled data of multiple first diabetics and the acquisition modes of fracture medical treatment result can
There are many, in one possible implementation, the sampled data and fracture medical treatment result of multiple first diabetics can
Think and obtained by obtaining multiple first diabetics in the second preset time period in the diagnosis records of multiple hospitals,
In, the second preset time period can be rule of thumb configured by those skilled in the art, and the second preset time period can be with
One preset time period is identical, or can also be different from the first preset time period, is specifically not construed as limiting.With multiple first diabetes
It may include blood sample achievement data, clinical indices in the sampled data of diabetic A for diabetic A in patient
At least one of in data, it can also include bone density achievement data and medical history achievement data, for example, diabetic A
It may include the blood sample achievement data and clinical indices data of diabetic A in sampled data, or may include diabetes
Blood sample achievement data, clinical indices data, bone density achievement data and the medical history achievement data of patient A, is specifically not construed as limiting.
Correspondingly, the fracture medical treatment result of diabetic A can be high fracture risk or fracture low-risk.
It should be noted that the sampled data of different diabetics involved in the embodiment of the present application may each be
The data obtained using identical index collection.In specific implementation, if the sampled data of a certain diabetic there are a certain item or
The shortage of data of certain indexs then can carry out completion processing to missing item.
Table 1 is the sampled data and the schematic table of fracture medical treatment result of a kind of multiple first diabetics, is shown in table 1
Multiple first diabetics sampled data and fracture medical treatment result can according to multiple first diabetics in 3 years
Diagnosis records obtain.
As shown in Figure 1, multiple first diabetics may include patient A, patient B, patient C, patient D, patient E, suffer from
Person F and patient G.By taking patient A as an example, the sampled data of patient A may include the blood sample achievement data of patient A, such as patient A
Albumin achievement data is 37.5g/L, and potassium achievement data is 4.6mmol/L, and total bilirubin achievement data is 1.8 μm of ol/L, blood
Amylase achievement data is 106.7U/L;Correspondingly, the sampled data of patient A may include the clinical indices data of patient A, than
If patient A weight be 66.2Kg, body temperature be 36.5 DEG C, pulse be 66 times/per minute.As shown in table 1, it can also be wrapped in table 1
The fracture medical treatment result of patient A is included, patient A is high risk fracture patient.
Step 202, at least one is obtained using the sampled data of multiple first diabetics and fracture medical treatment result training
A prediction model.
In the embodiment of the present invention, after the sampled data and fracture medical treatment result for getting multiple first diabetics,
Can the sampled data in advance to multiple first diabetics clean.To the sampled data of multiple first diabetics
The mode cleaned can there are many, be exemplified by Table 1, in one possible implementation, can be according to patient A~patient
The fracture medical treatment result of G rejects the patient F for not having fracture incidents in 3 years, it can by the sampled data of patient F and
Fracture medical treatment result is rejected from table 1;It is possible to further which missing data in patient A~patient E and patient G is more
Patient rejects, for example albumin, total bilirubin and the weight of patient B are missing item, therefore can be by the sampled data of patient B
It is rejected from table 1 with fracture medical treatment result.Correspondingly, the weight that the less patient D of item is lacked in table 1 can also be mended
Entirely, for example, the weight of patient D can be set to the middle position of the average value of the weight of other patients or the weight of other patients
Number.
It should be noted that above-mentioned is only a kind of illustrative simple declaration, cited by cleaning mode be only for
Convenient for illustrating scheme, does not constitute the restriction to scheme, in specific implementation, can also be cleaned using other way,
For example the index of the negligible amounts of the patient with a certain achievement data can be rejected, specifically it is not construed as limiting.By using upper
Mode is stated to clean the sampled data of multiple first diabetics, it can be based on multiple first patient of diabetes after cleaning
The sampled data and fracture medical treatment result training pattern of person, obtains at least one prediction model.
In the embodiment of the present invention, at least one prediction model may include Logic Regression Models, supporting vector machine model, with
Machine forest model and neural network model.The training process of each model is detailed below.
Logic Regression Models
It, can be by multiple first glycosurias after cleaning in specific implementation by taking Logic Regression Models are nonlinear model as an example
The sampled data and fracture medical treatment result of patient is divided into training data and verify data according to preset ratio, and can be preparatory
One initial non-linearities model is set, may include multiple unknown parameters in initial non-linearities model;By the way that training pattern is defeated
Enter and be trained in initial non-linearities model, can determine the value of multiple unknown parameters, to obtain the first nonlinear model.It lifts
For example, if the sampled data of multiple first diabetics after cleaning and fracture medical treatment result have 1000, preset ratio
For 4:1, then 800 in 1000 datas can be regard as training data, 200 datas in addition to 800 are as verifying
Data;Correspondingly, 800 training data training initial non-linearities models can be used, obtain the first nonlinear model.
It is possible to further use 200 verify datas to verify the first nonlinear model, in one example,
The parameter of the first nonlinear model can be optimized using the loss function and gradient descent method of maximal possibility estimation, be obtained
Target nonlinear model.Specifically, 20 verify datas are directed to, the sampled data that can will include in 20 verify datas
The first nonlinear model is inputted, prediction obtains the risk of bone fracture of 20 verify datas, further, if 20 bones that prediction obtains
Risk and 20 fracture medical treatment results are rolled over there are different, then the parameter that can slowly adjust in the first nonlinear model is (such as right
Parameter value successively+0.1, or successively -0.1), and the first nonlinear model that can be used after adjusting ginseng verifies 20 verifying numbers again
According to;Above-mentioned verification process is successively executed, until obtaining one group of target component, uses this group of target component that prediction can be made to obtain
20 risk of bone fracture and 20 fracture medical treatment results it is closest, then the corresponding nonlinear model of the target component be first object
Nonlinear model.
In one possible implementation, 1000 can repeatedly be drawn in such a way that preset ratio is 4:1
Point, multiple 800 training datas and 200 verify datas are obtained, and multiple 800 training datas and 200 can be used respectively
Verify data executes the process of above-mentioned model training, obtains multiple target nonlinear models;It is identical it is possible to further be based on
Verify data from multiple target nonlinear models determine optimal objective nonlinear model, the optimal objective nonlinear model was both
It can be the Logic Regression Models.
Supporting vector machine model
In specific implementation, can in advance the sampled data to multiple first diabetics after cleaning and fracture diagnosis and treatment knot
Fruit is normalized, wherein and the mode of normalized can be rule of thumb configured by those skilled in the art, than
Such as, each achievement data after can making normalized is any value in " 1 " and " 0 ", or may be in [0,1]
Arbitrary value, be specifically not construed as limiting.For example, being directed to Body Mass Index, in one example, can be set in 0~50kg
Body Mass Index data be 0, the Body Mass Index data greater than 50kg are 1;In another example, default letter can be set
Number (for example can be linear function, or may be nonlinear function), will be in 0~150kg by preset function
Weight is mapped as the respective value in [0,1].By the way that the data after cleaning are normalized, can eliminate due to absolute value
Weight bias problem caused by difference.
In the embodiment of the present invention, initial supporting vector machine model can be set, for example can be multinomial model, Huo Zheye
It can be Gauss model, or can also be Sigmoid function, specifically be not construed as limiting.Further, it is directed to normalized
Each data afterwards can convert mark sense for the achievement data of corresponding multiple first diabetics of an index
Amount, and a plurality of indicator vector after normalized can be divided into a plurality of training data and a plurality of verify data, using more
The initial supporting vector machine model of training data training, obtains the first supporting vector machine model;It is tested using a plurality of verify data
During demonstrate,proving the first supporting vector machine model, it can determine so that Optimal Parameters penalty factor and kernel functional parameter are optimal
Model parameter obtains the supporting vector machine model using the model parameter.
Random Forest model
In specific implementation, discretization can be carried out to the continuous achievement data in 1000 datas using dichotomy in advance,
Obtain the corresponding Liang Ge branch of each achievement data.For example, Body Mass Index can be divided into greater than 60kg, be less than or equal to
The Liang Ge branch of 60kg;For another example, height index can be divided into greater than 170cm, less than or equal to two points of 170cm
Branch;For another example, blood glucose target can be divided into greater than 5mmol/L, less than or equal to the Liang Ge branch of 5mmol/L.
It is possible to further which 1000 datas are divided into 800 training datas and 200 verify datas, and can be from
500 training datas are randomly selected in 800 training datas, by 500 training datas according to corresponding point of multiple pre-set levels
Zhi Jinhang modeling, obtains initial decision tree.During being verified using 200 verify datas to initial decision tree, root
It is predicted that obtained verification result and fracture medical treatment result are sequentially adjusted in the value of each score value and multiple branches in decision tree
Sequentially, objective decision tree is obtained.Fig. 3 is a kind of structural schematic diagram of initial decision tree.
As shown in figure 3, objective decision tree can be to be obtained based on the training of gender index, height index and albumin index
's.It is described by taking a prediction branch therein as an example, first branch of objective decision tree can be gender branch, gender
The branch value of corresponding first branch of branch can be to be male, and the branch value of the corresponding Article 2 branch of gender branch can be
Female;Second branch of the corresponding first branch road of gender branch can be height branch, and height branch is first corresponding
The branch value of branch can be for greater than 170 centimetres, the branch value of the corresponding Article 2 branch of height branch can be to be less than or wait
In 170 centimetres;The third branch of the corresponding first branch road of height branch can be albumin branch, albumin branch pair
The branch value for first branch answered can be for greater than 35mmol/L, and the corresponding prediction result of the branch is high fracture risk,
The branch value of the corresponding Article 2 branch of albumin branch can be for less than or equal to 35mmol/L, and the corresponding prediction of the branch
It as a result is fracture low-risk.If target diabetic's gender is male, height is 165 centimetres, albumin 40mmol/L, then root
According to objective decision tree illustrated in Figure 3, the prediction result of target diabetic can be determined for fracture low-risk.
In one possible implementation, 500 trained numbers can be repeatedly randomly selected from 800 training datas
According to, and can multiple objective decisions be obtained using 500 training datas repeatedly selected and 200 verify data training respectively
Tree, to obtain the Random Forest model based on the building of multiple objective decision trees.Correspondingly, if target mould in step 102
Type is Random Forest model, then each of Random Forest model objective decision tree can be used to target diabetic's
Risk of bone fracture is predicted, obtains multiple prediction results, and can be using a fairly large number of prediction result of prediction result as mesh
Mark the prediction result of diabetic.For example, including 100 objective decision trees in Random Forest model, wherein 80 targets are determined
Plan tree is high risk to the prediction result of the risk of bone fracture of target diabetic, and 20 objective decision trees are to target patient of diabetes
The prediction result of the risk of bone fracture of person is low-risk, then can determine that the prediction result of the risk of bone fracture of target diabetic is
High risk.
Neural network model
In specific implementation, the corresponding characteristic value of pre-set level can be obtained from 1000 datas in advance, pre-set level can
Think what those skilled in the art were rule of thumb configured, is specifically not construed as limiting.It is corresponding multiple based on each pre-set level
The index value of diabetic constitutes the corresponding feature vector of each pre-set level, and can be to the corresponding spy of multiple pre-set levels
Levy vector carry out dimension-reduction treatment, will feature vector corresponding with the fracture in patients with diabetes lower pre-set level of risk association degree into
Row is deleted, so as to obtain feature corresponding with the higher multiple pre-set levels of fracture in patients with diabetes risk association degree to
Amount.
It further, can be with after automatically generating multiple initial neural network models using automatic machinery learning method
Based on the initial neural network model of the corresponding feature vector training of multiple pre-set levels, it is with the first initial neural network model
Example, can optimize the connection relationship of each network layer, each net in the first initial neural network model by training
The quantity of active coating, pond layer, convolutional layer that network layers include, the value and quantity of the convolution kernel that each network layer includes, thus
So that activation primitive is optimal, i.e., so that the first initial neural network model training obtains the neural network model.Wherein,
Training objective neural network model process also may include using training data training pattern and using verify data Optimized model
Process, specific implementation process be referred to training obtains the implementation procedure in target logic regression model, repeat no more.
In the embodiment of the present invention, Logic Regression Models, supporting vector machine model, Random Forest model and neural network model
It is to be obtained by machine learning method training, on the one hand, predict diabetic's by using the method for machine learning
Risk of bone fracture can be realized with relatively advanced technology and be predicted the risk of bone fracture of patient of diabetes, to promote at data
Manage the development in medical domain;On the other hand, a variety of fracture in patients with diabetes risks are established by using the method for machine learning
Prediction model, the personalization that may be implemented to vary with each individual precisely diagnosis, for example, can be based on the sampling of target diabetic
Data determine object module, so that more accurate to the prediction result of target diabetic.
Step 203, object module is determined from least one prediction model.
In specific implementation, the sampled data and fracture medical treatment result of available multiple second diabetics, and will be more
The sampled data of a second diabetic inputs at least one prediction model, obtains the corresponding prediction result of each model.Its
In, multiple second diabetics be different from the patient other than above-mentioned described multiple first diabetics, for example,
The sampled data and fracture medical treatment result that 1200 diabetics are obtained in second preset time period, from 1200 patient of diabetes
1000 diabetics are selected in person as above-mentioned described first diabetic, and by remaining 200 glycosurias
Patient is as the second diabetic.
Further, each prediction model being directed at least one prediction model, can execute following process: will be more
The sampled data of a second diabetic inputs each prediction model and obtains the value-at-risk of multiple second fracture in patients with diabetes,
According to the fracture medical treatment result and multiple second diabetic bones of corresponding multiple second diabetics of each prediction model
The value-at-risk of folding determines the prediction effect of each prediction model.For example, table 2 is to use Logic Regression Models, supporting vector
The prediction effect that machine model, Random Forest model and neural network model predict 100 the second diabetics
Schematic table.
A kind of table 2: signal of the prediction effect of multiple prediction models
Prediction model | As a result correct | As a result mistake | Prediction effect |
Logic Regression Models | 950 | 50 | 95% |
Supporting vector machine model | 960 | 40 | 96% |
Random Forest model | 990 | 10 | 99% |
Neural network model | 850 | 150 | 85% |
As shown in table 2, the prediction effect of Logic Regression Models is 95%, and the prediction effect of supporting vector machine model is
96%, the prediction effect of Random Forest model is 99%, and the prediction effect of neural network model is 85%.Hence, it can be determined that
The prediction effect of Random Forest model is best, at this point it is possible to using Random Forest model as the object module in step 102.
It should be noted that above-mentioned is only a kind of illustrative simple declaration, cited by set the goal really the side of model
Formula is merely for convenience and purposes of illustration of scheme, does not constitute the restriction to scheme, and in specific implementation, object module can also be according to reality
Border needs to be configured, and is specifically not construed as limiting.In one example, can preset model each according to training when used finger
The corresponding relationship for the index for including in mark and the sampled data of target diabetic determines object module, for example, can will be with
The most prediction model of the corresponding index of the index for including in the sampled data of target diabetic is as object module.
In the above embodiment of the present invention, the sampled data of target diabetic is obtained, and by target diabetic
Sampled data input object module, obtain the value-at-risk of target fracture in patients with diabetes;Wherein, target diabetic adopts
Sample data include at least one in the blood sample achievement data and clinical indices data of target diabetic, and object module is root
It is obtained according to the sampled data and fracture medical treatment result training of multiple first diabetics.In the embodiment of the present invention, by making
Obtain object module with blood sample achievement data and the training of clinical indices data, may be implemented to the risk of bone fracture of diabetic into
Row prediction, and the prediction effect predicted the risk of bone fracture of diabetic can be made preferable.
For above method process, the embodiment of the present invention also provides a kind of device of determining fracture in patients with diabetes risk,
The particular content of the device is referred to above method implementation.
Fig. 4 is a kind of structural schematic diagram of the device of determining fracture in patients with diabetes risk provided in an embodiment of the present invention,
Include:
Module 401 is obtained, for obtaining the sampled data of target diabetic, the sampling of the target diabetic
Data include at least one in the blood sample achievement data and clinical indices data of the target diabetic;
Prediction module 402 obtains the mesh for the sampled data of the target diabetic to be inputted object module
Mark the value-at-risk of fracture in patients with diabetes;The object module is the sampled data and fracture according to multiple first diabetics
Medical treatment result training obtains.
Optionally, described device further includes training module, and the training module is used for:
It is pre- that at least one is obtained using the sampled data and fracture medical treatment result training of the multiple first diabetic
Survey model;
The each prediction model being directed at least one described prediction model, executes: by the multiple second diabetes
The sampled data of patient inputs the prediction model and obtains the value-at-risk of the multiple second fracture in patients with diabetes;According to described
Multiple second diabetics fracture medical treatment result and the multiple second fracture in patients with diabetes value-at-risk, determine described in
The prediction effect of prediction model;
Using the best prediction model of prediction effect at least one described prediction model as object module.
Optionally, at least one described prediction model includes Logic Regression Models, supporting vector machine model, random forest mould
Type and neural network model;
The training module is at least one of following for executing:
It is instructed using the sampled data and fracture medical treatment result of the multiple first diabetic by logistic regression algorithm
Get the Logic Regression Models;
Pass through algorithm of support vector machine using the sampled data and fracture medical treatment result of the multiple first diabetic
Training obtains the supporting vector machine model;
Multiple decision-tree models are obtained using the sampled data and fracture medical treatment result of the multiple first diabetic,
The Random Forest model is obtained according to the multiple decision-tree model;
Pass through neural network machine algorithm using the sampled data and fracture medical treatment result of the multiple first diabetic
Training obtains the neural network model.
It can be seen from the above: in the above embodiment of the present invention, the sampled data of target diabetic is obtained,
And the sampled data of target diabetic is inputted into object module, obtain the value-at-risk of target fracture in patients with diabetes;Wherein,
The sampled data of target diabetic include in the blood sample achievement data and clinical indices data of target diabetic extremely
One item missing, object module are obtained according to the sampled data and fracture medical treatment result training of multiple first diabetics.This
In inventive embodiments, object module is obtained by using blood sample achievement data and/or the training of clinical indices data, may be implemented pair
The risk of bone fracture of diabetic is predicted, and the prediction predicted the risk of bone fracture of diabetic can be made to imitate
Fruit is preferable.
Based on the same inventive concept, the embodiment of the invention also provides a kind of computer readable storage mediums, including instruct,
When run on a computer, so that computer executes the determination fracture in patients with diabetes wind as described in above-mentioned Fig. 1 embodiment
The method of danger.
Based on the same inventive concept, the embodiment of the invention also provides a kind of computer program products, when it is in computer
When upper operation, so that the method that computer executes the determination fracture in patients with diabetes risk as described in Fig. 1 embodiment.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method or computer program product.
Therefore, complete hardware embodiment, complete software embodiment or embodiment combining software and hardware aspects can be used in the present invention
Form.It is deposited moreover, the present invention can be used to can be used in the computer that one or more wherein includes computer usable program code
The shape for the computer program product implemented on storage media (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
Formula.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (11)
1. a kind of method of determining fracture in patients with diabetes risk, which is characterized in that the described method includes:
The sampled data of target diabetic is obtained, the sampled data of the target diabetic includes the target glycosuria
At least one of in the blood sample achievement data and clinical indices data of patient;
The sampled data of the target diabetic is inputted into object module, obtains the wind of the target fracture in patients with diabetes
Danger value;The object module is obtained according to the sampled data and fracture medical treatment result training of multiple first diabetics.
2. the method according to claim 1, wherein the object module is according to multiple first diabetics
Sampled data and fracture medical treatment result training obtain, comprising:
At least one prediction mould is obtained using the sampled data and fracture medical treatment result training of the multiple first diabetic
Type;
Obtain the sampled data and fracture medical treatment result of multiple second diabetics;
The each prediction model being directed at least one described prediction model, executes: by the multiple second diabetic
Sampled data input the prediction model and obtain the value-at-risk of the multiple second fracture in patients with diabetes;According to the multiple
The value-at-risk of the fracture medical treatment result and the multiple second fracture in patients with diabetes of second diabetic, determines the prediction
The prediction effect of model;
Using the best prediction model of prediction effect at least one described prediction model as object module.
3. according to the method described in claim 2, it is characterized in that, at least one described prediction model includes logistic regression mould
At least one of in type, supporting vector machine model, Random Forest model and neural network model;
At least one prediction mould is obtained using the sampled data and fracture medical treatment result training of the multiple first diabetic
Type includes at least one of the following:
Trained by logistic regression algorithm using the sampled data and fracture medical treatment result of the multiple first diabetic
To the Logic Regression Models;
Pass through algorithm of support vector machine training using the sampled data and fracture medical treatment result of the multiple first diabetic
Obtain the supporting vector machine model;
Multiple decision-tree models are obtained using the sampled data and fracture medical treatment result of the multiple first diabetic, according to
The multiple decision-tree model obtains the Random Forest model;
Pass through the training of neural network machine algorithm using the sampled data and fracture medical treatment result of the multiple first diabetic
Obtain the neural network model.
4. according to the method described in claim 1, it is characterized by:
The blood sample index includes fasting blood-glucose, postprandial 1 hour blood glucose, postprandial 2 hours blood glucose, glycosylated hemoglobins;It is described to face
Bed index includes gender, at the age, height, weight, smokes, drinks, family history.
5. according to the method described in claim 4, it is characterized in that, the blood sample index further includes following any one or any
It is multinomial:
Total cholesterol, triglycerides, aPoA, apolipoprotein B, high-density lipoprotein cholesterol, low-density lipoprotein gallbladder are solid
Alcohol.
6. the method according to any one of claims 1 to 5, which is characterized in that further include that bone is close in the sampled data
Spend achievement data and medical history achievement data;
The bone density index includes bone density, electronic constant computer tomography CT bone density;
The medical history index includes the past History of Bone Fracture, parent's Hip Fracture history.
7. a kind of device of determining fracture in patients with diabetes risk, which is characterized in that described device includes:
Module is obtained, for obtaining the sampled data of target diabetic, the sampled data bag of the target diabetic
Include at least one in the blood sample achievement data and clinical indices data of the target diabetic;
Prediction module obtains the target glycosuria for the sampled data of the target diabetic to be inputted object module
The value-at-risk of patient's fracture;The object module is the sampled data and fracture diagnosis and treatment knot according to multiple first diabetics
Fruit training obtains.
8. device according to claim 7, which is characterized in that described device further includes training module, the training module
For:
At least one prediction mould is obtained using the sampled data and fracture medical treatment result training of the multiple first diabetic
Type;
The each prediction model being directed at least one described prediction model, executes: by the multiple second diabetic
Sampled data input the prediction model and obtain the value-at-risk of the multiple second fracture in patients with diabetes;According to the multiple
The value-at-risk of the fracture medical treatment result and the multiple second fracture in patients with diabetes of second diabetic, determines the prediction
The prediction effect of model;
Using the best prediction model of prediction effect at least one described prediction model as object module.
9. device according to claim 8, which is characterized in that at least one described prediction model includes logistic regression mould
Type, supporting vector machine model, Random Forest model and neural network model;
The training module is used to execute at least one of following: using the sampled data and bone of the multiple first diabetic
Folding medical treatment result obtains the Logic Regression Models by the training of logistic regression algorithm;
Pass through algorithm of support vector machine training using the sampled data and fracture medical treatment result of the multiple first diabetic
Obtain the supporting vector machine model;
Multiple decision-tree models are obtained using the sampled data and fracture medical treatment result of the multiple first diabetic, according to
The multiple decision-tree model obtains the Random Forest model;
Pass through the training of neural network machine algorithm using the sampled data and fracture medical treatment result of the multiple first diabetic
Obtain the neural network model.
10. a kind of computer readable storage medium, which is characterized in that including instruction, when run on a computer, make to succeed in one's scheme
Calculation machine executes such as method as claimed in any one of claims 1 to 6.
11. a kind of computer program product, which is characterized in that when run on a computer, so that computer is executed as weighed
Benefit requires 1 to 6 described in any item methods.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910317945.9A CN110176304A (en) | 2019-04-19 | 2019-04-19 | A kind of method and device of determining fracture in patients with diabetes risk |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910317945.9A CN110176304A (en) | 2019-04-19 | 2019-04-19 | A kind of method and device of determining fracture in patients with diabetes risk |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110176304A true CN110176304A (en) | 2019-08-27 |
Family
ID=67689742
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910317945.9A Pending CN110176304A (en) | 2019-04-19 | 2019-04-19 | A kind of method and device of determining fracture in patients with diabetes risk |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110176304A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110853764A (en) * | 2019-11-28 | 2020-02-28 | 成都中医药大学 | Diabetes syndrome prediction system |
CN111009327A (en) * | 2019-12-19 | 2020-04-14 | 京东方科技集团股份有限公司 | Risk prediction method, device, system and medium |
CN111128327A (en) * | 2019-12-30 | 2020-05-08 | 中国科学院深圳先进技术研究院 | Hypoglycemia early warning method and device |
CN111599467A (en) * | 2020-05-26 | 2020-08-28 | 广州再生医学与健康广东省实验室 | Model training method and device, electronic equipment and medium |
CN111968748A (en) * | 2020-08-21 | 2020-11-20 | 南通大学 | Modeling method of diabetic complication prediction model |
CN115331825A (en) * | 2022-07-26 | 2022-11-11 | 上海交通大学医学院附属新华医院 | Fracture risk prediction model based on big data multi-level feature mining |
CN115662613A (en) * | 2022-09-28 | 2023-01-31 | 中日友好医院(中日友好临床医学研究所) | Barotrauma prediction method and device |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109065176A (en) * | 2018-08-20 | 2018-12-21 | 清华-伯克利深圳学院筹备办公室 | A kind of blood glucose prediction method, device, terminal and storage medium |
CN109308545A (en) * | 2018-08-21 | 2019-02-05 | 中国平安人寿保险股份有限公司 | The method, apparatus, computer equipment and storage medium of diabetes probability are suffered from prediction |
-
2019
- 2019-04-19 CN CN201910317945.9A patent/CN110176304A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109065176A (en) * | 2018-08-20 | 2018-12-21 | 清华-伯克利深圳学院筹备办公室 | A kind of blood glucose prediction method, device, terminal and storage medium |
CN109308545A (en) * | 2018-08-21 | 2019-02-05 | 中国平安人寿保险股份有限公司 | The method, apparatus, computer equipment and storage medium of diabetes probability are suffered from prediction |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110853764A (en) * | 2019-11-28 | 2020-02-28 | 成都中医药大学 | Diabetes syndrome prediction system |
CN110853764B (en) * | 2019-11-28 | 2023-11-14 | 成都中医药大学 | Diabetes syndrome prediction system |
CN111009327A (en) * | 2019-12-19 | 2020-04-14 | 京东方科技集团股份有限公司 | Risk prediction method, device, system and medium |
CN111128327A (en) * | 2019-12-30 | 2020-05-08 | 中国科学院深圳先进技术研究院 | Hypoglycemia early warning method and device |
CN111128327B (en) * | 2019-12-30 | 2023-05-26 | 中国科学院深圳先进技术研究院 | Hypoglycemia early warning method and device |
CN111599467A (en) * | 2020-05-26 | 2020-08-28 | 广州再生医学与健康广东省实验室 | Model training method and device, electronic equipment and medium |
CN111599467B (en) * | 2020-05-26 | 2023-12-05 | 生物岛实验室 | Model training method, device, electronic equipment and medium |
CN111968748A (en) * | 2020-08-21 | 2020-11-20 | 南通大学 | Modeling method of diabetic complication prediction model |
CN115331825A (en) * | 2022-07-26 | 2022-11-11 | 上海交通大学医学院附属新华医院 | Fracture risk prediction model based on big data multi-level feature mining |
CN115662613A (en) * | 2022-09-28 | 2023-01-31 | 中日友好医院(中日友好临床医学研究所) | Barotrauma prediction method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110176304A (en) | A kind of method and device of determining fracture in patients with diabetes risk | |
Sloan et al. | Vagally-mediated heart rate variability and indices of well-being: Results of a nationally representative study. | |
Vance et al. | The effects of physical activity and sedentary behavior on cognitive health in older adults | |
Voormolen et al. | Nonmaturation of arm arteriovenous fistulas for hemodialysis access: a systematic review of risk factors and results of early treatment | |
CN106175726A (en) | Human health status recognition methods based on BP neutral net | |
CN108475542A (en) | The method that remaining life is predicted by using biological age | |
JP6864947B2 (en) | How to create health positioning maps and health functions, systems, and programs, and how to use them | |
Cardoso et al. | The impact of frontal and cerebellar lesions on decision making: evidence from the Iowa Gambling Task | |
Dewan et al. | The role of frailty in failure to rescue after cardiovascular surgery | |
CN107194138A (en) | A kind of fasting blood-glucose Forecasting Methodology based on physical examination data modeling | |
Paliwal et al. | A model-based analysis of impulsivity using a slot-machine gambling paradigm | |
Freak-Poli et al. | Positive affect is not associated with incidence of cardiovascular disease: a population-based study of older persons | |
JP7398155B2 (en) | A system that estimates the health status of the human body | |
Vaidya et al. | The neuropsychology of decision-making: a view from the frontal lobes | |
Akman et al. | Determination of body fat percentage by gender based with photoplethysmography signal using machine learning algorithm | |
CN113178261A (en) | Diabetes prediction model construction method and system based on machine learning | |
Triantafillou | Against all odds? Understanding the emergence of accreditation of the Danish hospitals | |
Verharen et al. | Insensitivity to losses: A core feature in patients with anorexia nervosa? | |
CN109841278A (en) | A method of the screening of coronary disease disease is carried out with angiocarpy mark and rote learning operation method | |
Krueger et al. | Properties of the cognitive function battery for the MIND diet intervention to prevent Alzheimer’s disease | |
Johnson et al. | Propensity score analysis in observational studies: outcomes after abdominal aortic aneurysm repair | |
Persson et al. | Results from an intervention programme dealing with cardiovascular risk factors: Experiences from a study of men aged 33–42 in Habo, Sweden | |
Reanpang et al. | Upper limb exercise for people on haemodialysis following arteriovenous fistula surgery | |
CA3079087A1 (en) | Evaluating an individual's characteristics of at least one phenotype variable | |
Wang et al. | Chronic disease intervention in patients with cardiovascular disorders: a systematic review. |
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 |
Application publication date: 20190827 |
|
RJ01 | Rejection of invention patent application after publication |