CN103198211B - Quantitative analysis method for influences of attack risk factors of type 2 diabetes on blood sugar - Google Patents
Quantitative analysis method for influences of attack risk factors of type 2 diabetes on blood sugar Download PDFInfo
- Publication number
- CN103198211B CN103198211B CN201310074038.9A CN201310074038A CN103198211B CN 103198211 B CN103198211 B CN 103198211B CN 201310074038 A CN201310074038 A CN 201310074038A CN 103198211 B CN103198211 B CN 103198211B
- Authority
- CN
- China
- Prior art keywords
- crowd
- blood sugar
- hazards
- data
- output
- 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.)
- Expired - Fee Related
Links
Landscapes
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention relates to a quantitative analysis method for influences of attack risk factors of the type 2 diabetes on the blood sugar, and belongs to the fields of bioinformation processing and medicine. The method comprises the steps of using C4.5 and EM clustering algorithms to realize the selection of important attack risk factors firstly; then dividing all the crowd according to the gender and age; using a BP neural network algorithm to calculate the sensitivity of a specific crowd; and finally realizing the quantitative analysis of the influences of multiple factors on the blood sugar based on the sensitivity. Compared with a great quantity of conventional statistic methods, the invention adopts the data mining method, mutual influences among the multiple factors are considered fully, meanwhile, the quantitative analysis of the influences of the multiple factors on the blood sugar is realized in the specific crowd, the quantitative analysis accuracy is improved greatly, and a determination method can be provided for refining intervention of individual attack. The quantitative analysis method can be used for intervention instruction for the individual type 2 diabetes attack, not only can the attack be prevented or delayed, but also the method can be applied and popularized to the quantitative analysis of other disease risk factors.
Description
Technical field
The present invention relates to a kind of multifactor quantitative analysis method to blood sugar influence, belong to Bioinformatics and medical science neck
Domain.
Background technology
Diabetes B has become as a global major health concern.Expect 2025, the whole world will have 3.8
Hundred million people are perplexed by diabetes.At present, China has become diabetes second big country being only second to India.Aobvious according to Ministry of Public Health's investigation
Show, China diabetic about increases 3000 newly daily, about increase 1,200,000 newly every year, wherein about 95% is diabetes B patient.
Diabetes B has become after cancer and cardiovascular and cerebrovascular diseases, positioned at the 3rd chronic disease having a strong impact on human health, its disease
Because being the result of the interactions such as environmental factor, inherent cause, life style.The ill hazards of common recognition are obtained at present
Including increasing age, overweight, blood fat, blood pressure level exception, Diabetes family history etc., multifactor collective effect is to blood glucose level rises
High generation affects, and then leads to fall ill.
Due to diabetes B once morbidity is difficult to cure, if intervened to hazards in premorbid, can be effective
Reduce the incidence of disease, improve the quality of living.Correlative study adopts the statistical methods such as multiple regression, meta analysis, cox recurrence mostly,
Relation between studying hazards and whether fall ill using relative risk.The research of Harvard University Hu F B et al. shows to surpass
Weight and obesity are the most important factor that diabetes B occurs.Found by contrast, 3.4% is in low dangerous group women occurs glycosuria
The relative risk of disease is 0.09, and 91% patient that sends out is caused due to unhealthy habits and customs.Mhurchu C N et al. adopts
Reported with cox homing method and contact between the body mass index of Asian-Pacific area crowd and diabetes generation, find in this area's fall
Under-weight index can effectively reduce the incidence of disease of diabetes.Or adopting multivariate regression algorithm and meta analysis, research is generally with relatively
Risk factor illustrates whether a certain factor is the hazards occurring diabetes B related, provides qualitatively conclusion.The present invention adopts
BP neural network algorithm calculates susceptibility, quantifies to weigh the impact to change of blood sugar for the hazards, reflects danger by susceptibility
The impact to change of blood sugar for the change of dangerous factor, with the quantitative effect journey to change of blood sugar for the susceptibility comparative descriptions hazards
Degree, is that the process correlative factor to change of blood sugar feature and rule is explored, and for instructing corresponding intervening measure, controls blood sugar as early as possible
Rising trend, reach the pathogenetic purpose of prevention and control glycosuria.
Content of the invention
The purpose of the present invention be for solve the problems, such as multifactor to blood sugar influence quantitative analysis, propose a kind of based on BP nerve
The quantitative analysis method of network.
The design principle of the present invention is:Filter out main hazards using C4.5 and EM clustering algorithm, in order to determine
The object of quantitative analysis;To the national sample crowd's health check-up data not suffering from diabetes B, crowd is carried out according to sex and age
Divide;Using the impact to change of blood sugar for the BP neural network algorithm quantitative analysis hazards.The present invention filter out dangerous because
While plain, crowd is refined, by quantitative analysis multifactor to blood sugar influence, multifactor right in providing refinement crowd
Blood sugar influence quantization means, and the multifactorial quantization sequence difference of different refinement crowd, intervening for individual refinement provides judgement
Method.
The technical scheme is that and be achieved by the steps of:
Step 1, obtains crowd's health check-up data, forms national sample crowd's health check-up data source S not suffering from diabetes B.
Concrete grammar is:It is that health check-up data is surveyed by 2001-2008, obtain completely available data source, to health check-up
Data is pre-processed, and first passes through data scrubbing, fills vacancy value, identifies isolated point, eliminates noise and correct in data
Inconsistent;Carry out the conversion that data conversion includes Data Format Transform, data semantic again;Finally ensure the feelings do not lost in information
Under condition, by the factor that hough transformation deletion repeated factors and vacancy value are excessive, obtain national sample crowd's health check-up data source S=
{s1, s2, s3..., sk, wherein k is the sum of examinee after pretreatment.
Step 2, on the basis of step 1, carries out the screening of Major Risk Factors.Detailed process is as follows:
Step 2.1, data processing experiment parameter setting module.Select to carry out Major Risk Factors screening according to data source S
Algorithm, and the parameter of set algorithm.
Step 2.2, EM cluster arithmetic module.
Concrete grammar is:Data source S is carried out with the cluster experiment of poly- P class or q class, changes the hazards participating in experiment
Value volume and range of product, observation experiment result, obtain preferably reflecting the cluster result of group characteristic, record participates in the danger of cluster
Dangerous factor.
Step 2.3, EM cluster, the experiment of C4.5 sort merge.
Concrete grammar is:The participation factor that EM clusters experimental section is the optimal cluster factor that above-mentioned cluster tests gained,
Carry out the cluster experiment of poly- P class or q class, data source S is pressed different crowd health feature separately, in the people to different health features
Group is analyzed using C4.5 algorithm respectively, and classification participation factor is that whole l tie up hazards, the demarcation threshold value of classification experiments
It is respectively R, V, T and Z, obtain the categorised decision tree corresponding to different health feature crowds.
Step 2.4, counts to experimental result, obtains c dimension Major Risk Factors, cognitive according to medical science, sieves further
Choosing obtains u dimension Major Risk Factors.Step 3, according to sex and age, to the national sample crowd's health check-up number obtaining through step 2
Divided according to source S, generated refinement crowd.
Concrete grammar is:Press sex first to divide, obtain males and women population;Be more than respectively by age again e year
Divided with less than or equal to e year, d group refinement crowd is obtained.
Step 4, is respectively trained BP neural network model using the refinement crowd obtaining through step 3, and then calculates difference
The susceptibility to blood sugar influence for the hazards, realizes quantitative analysis using susceptibility.
Step 4.1, under given Major Risk Factors dimension u, is trained using d group refinement crowd and generates d BP nerve net
Network model, the generation method of each model is:
Step 4.1.1, the u dimension hazards of training data after selection process, as the input of model, blood sugar is as model
Output, the backpropagation training using the forward-propagating of information and error generates BP neural network model.Input hazards
Successively calculate through hidden layer from input layer and be delivered to output layer, each layer of neuron only affects the state of next layer of neuron, such as
Fruit output layer does not obtain desired output, then calculate the error change value of output layer, then carry out backpropagation, will by network
Error signal returns to adjust the weights of each neuron along original connecting path anti-pass, through successive ignition, until it reaches average
Relative error is less than σ, and training generates BP neural network model, and computation model exports average relative error.
Step 4.1.2, then the BP neural network model checking data input generation, calculate output blood glucose value, pass through
Error calculation is verified the average relative error of data.
Step 4.2, calculates the multifactor susceptibility to blood sugar influence by BP neural network model.Susceptibility is by dividing
The impact to modeling effect for the analysis different parameters combination, the contribution rate that the model parameter determined exports on model or impact journey
Degree.
(n is the number of BP neural network mode input variable, and L is BP neural network model to be provided with n-L-1 feedforward network
Implicit number of layers, 1 is the number of model output variable), network output has following form:y=f(x1..., xn)(X is BP nerve net
The input of network model, y is the output of BP neural network model).Taking 2 input hazards as a example, by second order is asked to this formula
Local derviation is investigating the susceptibility to output variable for two input variables.If the hidden layer activation primitive of neutral net is logarithm S type letter
Number
By Jacobian matrix
In formula:T is the transposition computing of matrix, and m is the number of samples of data source used, and n is the number of input variable.?
J input xjChange and j-th output yj=f(xj) change connect mean network output susceptibility depend on input
Small sample perturbations.For n input, there is the hidden layer of L neuron and the neutral net of an output layer, on t-th sample
Input variable xiAnd xkTo the susceptibility of output variable y it is
In formula:S1The first derivative it being inputted for output layer activation primitive, S2For output layer activation primitive, it is inputted
Second dervative.For the response of j-th hidden neuron on t-th sample, vj1For output neuron and j-th hidden neuron
Between weight, wijFor the weight between i-th input neuron and j-th hidden neuron, wkjFor k-th input neuron and
Weight between j-th hidden neuron.By sensitivity analysis is carried out to different hazards, obtain each risk factors pair
The quantitative analysis of change of blood sugar.
Beneficial effect
Compared to based on a large amount of statistical analysis method such as linear regression, meta analysis, the present invention adopts BP neural network
Data digging method, realizes the quantitative analysis to change of blood sugar, has the characteristics that accuracy rate is high.
Compared with population analysis, the present invention adopts crowd's partitioning technology, has higher accuracy rate, change of blood sugar is divided
Analysis is more targeted, and provides judgment basis for individual refinement intervention, to prevent or to delay the generation of diabetes B.This
The bright quantitative analysis being applied and popularized to Other diseases hazards, applies also for the good of factor intervention-judgement-factor intervention
Property circulation in, thus effectively lifted individuality the general level of the health.
Brief description
Fig. 1 is the multifactor schematic diagram to blood sugar influence quantitative analysis method of the present invention;
Fig. 2 is data prediction schematic diagram in specific embodiment;
Fig. 3 is to cluster experiment flow figure in specific embodiment;
Fig. 4 is cluster, sort merge experiment flow figure in specific embodiment;
Fig. 5 is crowd's division methods in specific embodiment;
Fig. 6 is BP neural network model generating principle figure in specific embodiment;
Fig. 7 is risk factors susceptibility histogram in specific embodiment.
Specific embodiment
In order to better illustrate objects and advantages of the present invention, below in conjunction with the accompanying drawings with embodiment to the inventive method
Embodiment is described in further details.
Using MEC 2007 and 9632 actual measurement health check-up data in 2008 as input, design and dispose 9 groups of refinements
The checking of crowd:(1) verified for whole demographic data, (2) are verified for male sex's data, (3) are directed to women number
According to being verified, (4) are directed to and are verified more than 50 years old demographic data, and (5 are directed to less than or equal to demographic data checking in 50 years old, (6)
It is more than demographic data checking in 50 years old for the male sex, (7) are directed to the male sex and are less than or equal to demographic data checking in 50 years old, and (8) are directed to women
Verify more than 50 years old demographic data, (9 are directed to women is less than or equal to demographic data checking in 50 years old.
Training data is derived from 2001-2008 and surveys health check-up data, has 59839 and is not suffering from sick body inspection data as defeated
Enter, wherein male sex's data 34377, account for 57.4%, women data 25462, account for 42.6% whether with diabetes B according to
The World Health Organization in 1999(WHO)Standard determination.The concrete sex of examinee and age distribution are as shown in table 1.Checking data
Using MEC 2007 and 200 years 9632 actual measurement health check-up data.
The sex of table 1 training data and age distribution statistical form
Training data is carried out with data prediction such as Fig. 2 and crowd divides such as Fig. 3, obtain nine groups of crowds, be respectively trained life
Become nine BP neural network model such as Fig. 4, calculate average relative error:
In formula:The average relative error that E exports for model, m is the number of samples of data source used, y'iFor model output
I-th sample blood glucose value, yiFor the actual blood glucose value of i-th sample, it is calculated the average relative of each model respectively by mistake
Difference.
Same data prediction is carried out to checking data and crowd divides, obtain nine groups of crowds, input corresponding mould respectively
Type, calculates each model and exports blood glucose value, then pass through error calculation average relative error, to verify the accuracy of model.
The average relative error table of table 2 nine group model output
Respectively obtain nine groups of crowd's illness hazards susceptibilitys to change of blood sugar by nine BP neural network models
As shown in table 3, corresponding risk factors susceptibility histogram is as shown in figure 5, be from left to right followed successively by entire population, the male sex
Crowd, women population, more than 50 years old crowd, be less than or equal to 50 years old crowd, the male sex>50 years old crowds, women>50 years old crowds, the male sex
The Bu Tong ill hazards susceptibility of≤50 years old crowds and women≤50 year old crowd.
3 nine groups of crowd's illness hazards susceptibility tables to change of blood sugar of table
By the susceptibility comparative analysis of hazards in different crowd it can be deduced that following result:
1. body weight affects on change of blood sugar
Body weight is the factor of easily attractive change of blood sugar.The susceptibility that in full crowd, body weight affects on change of blood sugar is
0.2449.Body weight the influence degree of change of blood sugar is not only embodied in full crowd susceptibility calculate in body weight susceptibility rank the
One;And be presented as after application sex, the age is further divided into 8 crowds, wherein in 6 crowds body weight susceptibility
It is respectively positioned on first, be more than with the age in 50 years old women population only in crowd more than 50 years old and be located at second and three respectively.And
Afterwards two groups be also contemplated for being all the feature that the age is more than 50 years old women population group caused by.
2. the impact to change of blood sugar for the blood lipid level
The impact to change of blood sugar for the cholesterol levels change is only second to body weight.Cholesterol levels change sensitivity in full crowd
For 0.2294.At the age more than 50 years old and in women population, cholesterol levels are higher than the age less than 50 years old group to blood sugar influence
And the male sex.It is more than in the women population of 50 years old at the age, the impact to change of blood sugar for the cholesterol levels is higher than same age group man
The 28% of property crowd(0.2538vs.0.1985).In full crowd, triglyceride levels are located at the 4th to change of blood sugar impact
(0.1227), low by 47% far beyond the impact of cholesterol levels(0.2294vs.0.1227).But in males, particularly age
More than in 50 years old males, triglyceride levels affect on change of blood sugar(0.1970)Substantially increase by 60%.
3. the impact to change of blood sugar for the age
In full crowd, the impact susceptibility to change of blood sugar for the age is 0.1657, positioned at the 3rd.The male sex is than women couple
Age factor will sensitivity(0.2192vs.0.0383)4.7 again.
In full crowd, susceptibility that three above factor affects on change of blood sugar reached current detection ill dangerous because
The 64% of element impact.If it is considered that triglycerides is 0.1227 to the susceptibility of change of blood sugar, body weight, blood lipid level(Cholesterol
And triglycerides)Can be to 77% with the impact to change of blood sugar for three factors of age, three respectively may be about 25%, 35% and 17%.
In general, body weight, blood lipid level(Cholesterol and triglycerides)It is the principal element of impact change of blood sugar with age factor.
4. sex affects on change of blood sugar
In full crowd, the susceptibility of sex is only 0.0091, and the impact to change of blood sugar for the sex factor is in full crowd
Less, this also complies with the little phenomenon of diabetes prevalence gap between sex for effect.But after considering age factor, sex
Impact to change of blood sugar still has certain effect.What in more than 50 years old crowd, sex affected on blood sugar level improves near
2.5 times, susceptibility brings up to 0.0315.
In age groups, sex factor shows as on the concrete impact of change of blood sugar:
1)More than in 50 years old masculinity and femininity group, blood lipid level change impact embodies different.The male sex is for glycerine
Three ester level changes are more sensitive(0.1970vs.0.1659), and cholesterol and hdl level change for women
Blood sugar influence is bigger(0.2538vs.0.1985;0.1974vs.0.1437).
2)Less than in 50 years old male sex's group, body weight factor shows highest influence degree(0.2911), exceed complete respectively
People group, simple male sex's group and the male sex are higher than the 19%, 18% and 32% of 50 years old group influence degree.Cholesterol and HDL
For women blood sugar influence still existing in women group less than 50 years old, and sex factor is in high density lipoprotein level plain boiled water for level change
Impact on flat became apparent from more than 50 years old group degree than the age(0.1515vs.0.0371).
Obtain further as drawn a conclusion:Based on the quantitative analysis method of BP neural network, weigh danger using sensitive metrization
Dangerous factor is to the influence degree of change of blood sugar it is achieved that from qualitative analysis to the quantitative transformation calculating, having obtained different ill danger
The susceptibility that dangerous factor affects on change of blood sugar.(1) changes of weight easily attractive change of blood sugar, next to that cholesterol, the age and
Triglycerides, they reach the 77% of the ill hazards of current detection to the susceptibility of change of blood sugar impact, body weight, blood
Lipid level(Cholesterol and triglycerides)Account for 25%, 35% and 17% with the age respectively.(2)The impact to change of blood sugar for the age is quick
Sensitivity is 0.1657, positioned at the 3rd.The male sex is more sensitive to age factor than women(0.2192vs.0.0383)4.8 again.(3)
The impact to change of blood sugar for the sex factor acts on less in full crowd, and susceptibility is 0.0091, but after considering age factor, property
Other have certain effect to change of blood sugar.In more than 50 years old crowd, more substantially, susceptibility is for the impact to blood sugar level for the sex
0.0315.More than in 50 years old crowd, the impact to blood sugar level for male sex's triglycerides is higher by 19% than women;Body weight is to male sex's blood sugar
The impact of level is higher by 14% than women, and the impact to change of blood sugar for the women cholesterol levels is higher by 28% than the male sex;Less than or equal to 50 years old
In crowd, HDL is obvious to the protective effect of change of blood sugar.
Claims (3)
1.2 patients with type Ⅰ DM risk factors to the quantitative analysis method of blood sugar influence it is characterised in that methods described include with
Lower step:
Step 1, surveys health check-up data to 2001-2008, carries out data scrubbing, and filling vacancy value, identification isolated point, elimination are made an uproar
Sound is simultaneously corrected inconsistent in data;Carry out data conversion again, including Data Format Transform, data semantic conversion;Finally protecting
In the case that card information is not lost, repeated factors are deleted by data regularization and vacancy is worth more factor, formed and do not suffer from 2 types
National sample crowd's health check-up data source S of diabetes;
Step 2, carries out hazards to data source S using EM clustering algorithm and roughly selects, then using fusion EM cluster and C4.5
The hazards method for concentrating of algorithm, screening causes the Major Risk Factors of diabetes B;
Step 3, according to sex and age, divides to national sample crowd's health check-up data source S obtaining through step 1, is based on
Step 2 obtains hazards and is respectively trained BP neural network model to 9 groups of refinement crowds, based on BP neural network weight, adopts
A kind of susceptibility computational methods under multifactor functioning, calculate the susceptibility to blood sugar influence for the different hazards, enter
And realize quantitative analysis;
Wherein, the susceptibility computational methods under described multifactor functioning are:It is provided with n-L-1 feedforward network, in formula, n is BP god
Number through network model input variable, L is the implicit number of layers of BP neural network model, and 1 is the number of model output variable,
Network output has following form:Y=f (x1..., xn), in formula, x is the input of BP neural network model, and y is BP neural network mould
The output of type, investigates the susceptibility to output variable for the input variable by this formula is asked with second order local derviation, if neutral net is hidden
Layer activation primitive is logarithm S type function
By Jacobian matrix
In formula:T is the transposition computing of matrix, and m is the number of samples of data source used, and n is the number of input variable, j-th
Input xjChange and j-th output yj=f (xj) change connect mean network output susceptibility depend on the micro- of input
Microvariations, for n input, have the hidden layer of L neuron and the neutral net of an output layer, defeated on t-th sample
Enter variable xiAnd xkTo the susceptibility of output variable y it is
In formula:S1The first derivative it being inputted for output layer activation primitive, S2The second order it being inputted for output layer activation primitive
Derivative,For the response of j-th hidden neuron on t-th sample, vj1For between output neuron and j-th hidden neuron
Weight, wijFor the weight between i-th input neuron and j-th hidden neuron, wkjFor k-th input neuron and j-th
Weight between hidden neuron.
2. method according to claim 1 is it is characterised in that described carry out danger to data source S using EM clustering algorithm
Roughly selecting of factor, then using the hazards method for concentrating merging EM cluster and C4.5 algorithm, screens and causes diabetes B
The step of Major Risk Factors specifically includes:
Step 2.1, selects to carry out the algorithm of Major Risk Factors screening according to data source S, and the parameter of set algorithm;
Step 2.2, carries out the cluster experiment of poly- P class or q class to data source S, change the quantity of hazards participating in experiment and
Species, observation experiment result, obtain preferably reflecting the cluster result of group characteristic, record participate in the danger of cluster because
Element, reaches the purpose that hazards are roughly selected;
Step 2.3, the participation factor that EM clusters experimental section is the optimal cluster factor that above-mentioned cluster tests gained, carries out poly- P
The cluster experiment of class or q class, data source S is pressed different crowd health feature separately, in the crowd to different health features respectively
It is analyzed using C4.5 algorithm, classification participation factor is that whole l tie up hazards, and the demarcation threshold value of classification experiments is respectively
A, B, C and D, obtain the categorised decision tree corresponding to different health feature crowds;
Step 2.4, counts to experimental result, obtains c dimension Major Risk Factors, cognitive according to medical science, screens further
Tie up Major Risk Factors to u, reach the purpose selected to hazards.
3. method according to claim 1 is it is characterised in that according to age-sex, to the national sample obtaining through step 1
Crowd's health check-up data source S is divided, and is respectively trained BP neural network model to refinement crowd, based on BP neural network weights meter
Calculate the susceptibility to blood sugar influence for the different hazards, and then realize quantitative analysis, concrete grammar is:
Step 3.1, according to age and sex, to step 1, crowd refines, first press sex divide, obtain males and
Women population;Again respectively by age be more than e year and less than or equal to e year divided, d group refinement crowd is obtained;
Step 3.2, under given Major Risk Factors dimension u, is trained using n group refinement crowd and generates n BP neural network mould
Type, the generation method of each model is:
Step 3.2.1, the u dimension hazards of training data after selection process, as the input of model, blood sugar is defeated as model
Go out, generate BP neural network model using the forward-propagating of information and the backpropagation training of error, input hazards are from defeated
Enter layer and successively calculate through hidden layer to be delivered to output layer, each layer of neuron only affects the state of next layer of neuron, if defeated
Go out layer and do not obtain desired output, then calculate the error change value of output layer, then carry out backpropagation, by network by error
Signal returns to adjust the weights of each neuron along original connecting path anti-pass, through successive ignition, until it reaches average relative
Error is less than σ, and training generates BP neural network model, and computation model exports average relative error;
Step 3.2.2, then the BP neural network model checking data input generation, calculate output blood glucose value, by error
It is calculated the average relative error of checking data;
Step 3.3, susceptibility is the impact by analyzing different parameters combination to modeling effect, the model parameter determined
Contribution rate to model output or influence degree, by carrying out sensitivity analysis to different hazards, obtain each initiation potential
The quantitative analysis results to change of blood sugar for the factor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310074038.9A CN103198211B (en) | 2013-03-08 | 2013-03-08 | Quantitative analysis method for influences of attack risk factors of type 2 diabetes on blood sugar |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310074038.9A CN103198211B (en) | 2013-03-08 | 2013-03-08 | Quantitative analysis method for influences of attack risk factors of type 2 diabetes on blood sugar |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103198211A CN103198211A (en) | 2013-07-10 |
CN103198211B true CN103198211B (en) | 2017-02-22 |
Family
ID=48720765
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310074038.9A Expired - Fee Related CN103198211B (en) | 2013-03-08 | 2013-03-08 | Quantitative analysis method for influences of attack risk factors of type 2 diabetes on blood sugar |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103198211B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107194138B (en) * | 2016-01-31 | 2023-05-16 | 北京万灵盘古科技有限公司 | Fasting blood glucose prediction method based on physical examination data modeling |
CN107491656B (en) * | 2017-09-04 | 2020-01-14 | 北京航空航天大学 | Pregnancy outcome influence factor evaluation method based on relative risk decision tree model |
CN107742533A (en) * | 2017-09-13 | 2018-02-27 | 天津大学 | A kind of method based on dietary nutrient prediction gout |
TWI678630B (en) * | 2018-01-16 | 2019-12-01 | 中央研究院 | A system and method for non-invasively estimating hba1c and blood glucose level |
CN108877943A (en) * | 2018-06-21 | 2018-11-23 | 天津医科大学 | Type-2 diabetes mellitus risk evaluation model based on evidence-based medical |
CN109360604B (en) * | 2018-11-21 | 2021-09-24 | 南昌大学 | Ovarian cancer molecular typing prediction system |
CN109686442B (en) * | 2018-12-25 | 2020-04-14 | 南京市中西医结合医院 | Machine learning-based gastroesophageal reflux disease risk factor determination method and system |
CN110575181A (en) * | 2019-09-10 | 2019-12-17 | 重庆大学 | Near infrared spectrum noninvasive blood glucose detection network model training method |
CN111816310A (en) * | 2020-07-16 | 2020-10-23 | 山东大学 | Bone marrow blood disease risk factor contribution rate calculation and risk prediction system |
CN113270194A (en) * | 2021-04-22 | 2021-08-17 | 深圳市雅士长华智能科技有限公司 | Health data management system based on cloud computing |
CN118039168A (en) * | 2024-04-11 | 2024-05-14 | 陕西爱普医疗器械有限公司 | Diabetes risk prediction method |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102930163A (en) * | 2012-11-01 | 2013-02-13 | 北京理工大学 | Method for judging 2 type diabetes mellitus risk state |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7392140B2 (en) * | 2003-09-23 | 2008-06-24 | Prediction Sciences, Llc | Cellular fibronectin as a diagnostic marker in stroke and methods of use thereof |
US7634360B2 (en) * | 2003-09-23 | 2009-12-15 | Prediction Sciences, LL | Cellular fibronectin as a diagnostic marker in stroke and methods of use thereof |
WO2008131224A2 (en) * | 2007-04-18 | 2008-10-30 | Tethys Bioscience, Inc. | Diabetes-related biomarkers and methods of use thereof |
-
2013
- 2013-03-08 CN CN201310074038.9A patent/CN103198211B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102930163A (en) * | 2012-11-01 | 2013-02-13 | 北京理工大学 | Method for judging 2 type diabetes mellitus risk state |
Non-Patent Citations (2)
Title |
---|
《多维2型糖尿病实测数据的预处理技术》;罗森林 等;《计算机工程》;20040930;第30卷(第17期);摘要,第1.2节 * |
《改进的BP神经网络在糖尿病危险因素分析中的应用》;王美,刘泱;《软件导刊》;20080430;第7卷(第4期);摘要,第2节,表1,表3 * |
Also Published As
Publication number | Publication date |
---|---|
CN103198211A (en) | 2013-07-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103198211B (en) | Quantitative analysis method for influences of attack risk factors of type 2 diabetes on blood sugar | |
CN109119167B (en) | Sepsis mortality prediction system based on integrated model | |
CN109528197B (en) | Individual prediction method and system for mental diseases based on brain function map | |
Ahmed et al. | Single volume image generator and deep learning-based ASD classification | |
Pieroth et al. | The relationship between social support and diet quality in middle-aged and older adults in the United States | |
Schiele et al. | Quantile treatment effects of job loss on health | |
CN103714261B (en) | Intelligent auxiliary medical treatment decision supporting method of two-stage mixed model | |
CN102930163A (en) | Method for judging 2 type diabetes mellitus risk state | |
CN107463766A (en) | Generation method, device and the computer-readable recording medium of blood glucose prediction model | |
CN106446566A (en) | Elderly cognitive function classification method based on random forest | |
CN108511056A (en) | Therapeutic scheme based on patients with cerebral apoplexy similarity analysis recommends method and system | |
CN109106384B (en) | Psychological stress condition prediction method and system | |
CN114999629A (en) | AD early prediction method, system and device based on multi-feature fusion | |
CN112472048A (en) | Neural network structure for pulse condition recognition of cardiovascular disease patient | |
CN111105877A (en) | Chronic disease accurate intervention method and system based on deep belief network | |
CN115602325A (en) | Chronic disease risk assessment method and system based on multi-model algorithm | |
CN105868532A (en) | Method and system for intelligently evaluating heart ageing degree | |
CN114203295A (en) | Cerebral apoplexy risk prediction intervention method and system | |
CN114943629A (en) | Health management and health care service system and health management method thereof | |
Wang et al. | Simultaneous prediction of peach firmness and weight using vibration spectra combined with one-dimensional convolutional neural network | |
CN114300126A (en) | Cancer prediction system based on early cancer screening questionnaire and feed-forward neural network | |
CN112967807B (en) | System, device and storage medium for predicting cerebral apoplexy based on diet behavior | |
CN111916179A (en) | Method for carrying out 'customized' diet nourishing model based on artificial intelligence self-adaption individual physical sign | |
CN111445980B (en) | Meal scheme reactivity prediction model, modeling method and electronic equipment | |
CN108346468B (en) | Data processing method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170222 Termination date: 20200308 |