CN109920547A - A kind of diabetes prediction model construction method based on electronic health record data mining - Google Patents
A kind of diabetes prediction model construction method based on electronic health record data mining Download PDFInfo
- Publication number
- CN109920547A CN109920547A CN201910161929.5A CN201910161929A CN109920547A CN 109920547 A CN109920547 A CN 109920547A CN 201910161929 A CN201910161929 A CN 201910161929A CN 109920547 A CN109920547 A CN 109920547A
- Authority
- CN
- China
- Prior art keywords
- data
- electronic health
- health record
- diabetes
- prediction model
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of diabetes prediction model construction method based on electronic health record data mining, this method is from the building of electronic health record data cleansing and prediction model, the electronic health record data come from the export of each server are carried out to the integration of data by unique patient's identification number, data include essential information and diagnosis, saccharification and biochemical analysis data, by essential information, diagnostic message etc. is merged into a complete sample.The cleaning of data is carried out to data, the data after cleaning are stored in database by removal abnormal data, repeated data and existing wrong data.Prediction is classified to the diabetes data after cleaning, it can be concluded that, no matter nicety of grading or model-evaluation index are better than other algorithm models to improved BP-NN model by result.This method improves the recall rate for not examining nephrosis in crowd, enhances nephrosis control efficiency, and save a large amount of health resources.
Description
Technical field
The invention belongs to the chronic disease risk profile fields that medical big data is excavated, and are related to a kind of method of model construction,
Especially with regard to a kind of chronic disease risk forecast model construction method based on electronic health record data mining.
Background technique
Diabetes are to endanger a kind of more universal chronic disease of health of people life instantly.Diabetes have become after heart and brain
The chronic disease of the third-largest threat human health after vascular diseases, malignant tumour.With generally mentioning for our people's living standard
High and the accelerating rhythm of life, patient of diabetes patient quantity increase at an amazing speed, and develop to becoming younger.It is newest
Investigation display, China maturity-onset diabetes patient are up to 1.14 hundred million, are presented high incidence state, however awareness, treatment rate and reach
Mark rate is relatively low.Not getting timely medical treatment may have cardiovascular and cerebrovascular and diabetes etc. various simultaneously with the diabetic of control
Disease is sent out, this has not only seriously affected the quality of life of diabetic, has also brought heavy burden for family and society.Cause
This, the generation of prevention and control diabetes for save medical resources, reduces China's medical expense with own strategic significance.
With the arrival of information age, information-based tide equally promotes the development of hospital information, electronic health record at
For the chief component of Hospital Information Systems, the application of present electronic health record also becomes a kind of trend.Hospital
There are whole clinical datas of the diagnosis and treatment process of a large amount of diabetic in electronic health record, mainly include that electronic health record is believed substantially
Cease medical informations such as (name, gender, occupation etc.), medication information, Indexs measure data, doctor's advice etc..It is existing that these are a large amount of
Information in more under cover with the knowledge of diabetic complication occurrence regularity and diagnosis and treatment means.Pass through electronic health record data mining
Risk is made prediction, these hiding rules excavated are made good use of, and auxiliary doctor carries out the prediction of diabetic complication,
So that doctor is directed to complication risk, relevant treatment is carried out ahead of time, controls the development of the state of an illness, further decreases diabetes in people
Disease incidence in group, finally substantially reduces the harm of diabetes, this is a critically important help to the diagnosis and treatment of diabetes.
There is presently no the methods of radical cure diabetes, and some treatment means are usually used to control the development of the state of an illness.But
It is complication caused by diabetes is to endanger health of people to even result in four killers often, diabetic complication has following two
A feature, first is that complication early stage is not easy to find, and current treatment of diabetes is still based on the diagnosis and treatment of doctor, practitioner
The case where according to the performance of the symptom of patient, every body index, historical therapeutic situation etc. combine the experience of doctor to accumulate
The diagnosis and treatment method of electronic health record out, and since the personal feature of electronic health record has differences, the similar state of an illness uses identical
Often result is also far different to diagnosis and treatment method；Second is that complication is once generate, drug therapy is difficult to reverse, and caused by result
Be it is very serious, even result in death.Therefore it emphasizes to prevent diabetic complication as early as possible, early discovery and early treatment.
What the domestic method using data mining was multi-purpose greatly to the early stage research of diseases analysis at present is that mature data are dug
Dig software such as (SPSS, SAS etc.), in conjunction with medical data obtained to given data attribute set carry out onset risk and
The analysis of factor.There are also some scholars using machine learning methods such as post-class processing, support vector machines to electronic health record data
It carries out the foundation of model and obtains corresponding risk factor.External scholar carries out the prediction of chronic disease using data mining technology
It starts to walk more early, it is relatively mature in method, there are many method and model to be applied to research process.They compare difference
Feature selecting algorithm and different machine learning algorithms studied to the prediction effects of diabetes, and using the correlation rule expanded
To diabetes and relevant factor and be susceptible to suffer from diabetic population.
In recent years, deepening continuously with the research of deep learning, artificial intelligence in medical industry using more and more,
It is also following big potential developing direction.The essence of artificial intelligence is neural network, the prediction model of this building
Also neural network will be introduced and be trained prediction.The neural network number of plies can be chosen according to sample size, once sample size data
It is too small, cause identification error rate to greatly promote.The clinic diagnosis data of diabetes can be related to the data of many dimensions and very much
Field attribute the characteristics of for obtained data, has chosen the prediction model based on BP neural network, while for number
According to irregular timing feature the algorithm is improved, can dog preferably reach the result of prediction.
In conclusion being established by the processing to diabetes electronic health record data, and with improved neural network to data
Prediction model improves the accuracy rate of the clinical decision of diabetes, and can implement the early intervention of diabetes, and diabetes are suffered from reduction
And suffer from the risk of diabetic complication.
Summary of the invention
Current diabetes there are aiming at the problem that, the object of the present invention is to provide a kind of accurate bases reasonable, easy to use
In the diabetes illness risk forecast model construction method of electronic health record data mining.
To achieve the above object, the present invention takes following technical scheme: a kind of glycosuria based on electronic health record data mining
Disease forecasting model building method, steps are as follows:
Step 1: the electronic health record data come from the export of each server are subjected to the whole of data by unique patient's identification number
It closes, the essential information of electronic health record and diagnostic message etc. is merged into a complete sample.
Step 2: by electronic health record data prediction, obtaining clean available electronic health record data and be stored in database.
Step 3: analyzing resulting electronic health record data cases, several risk factors of diabetes are determined, according to glycosuria
The characteristics of sick data, improves the BP neural network of selection, and establishes diabetes prediction model on this basis.According to sugar
The characteristics of urinating sick data improves the BP neural network of selection, and establishes diabetes prediction model on this basis, and make
It is tested with processed electronic health record data.
Step 4: model is carried out to diabetes data using k nearest neighbor, logistic regression, decision tree, random forests algorithm etc.
Training prediction, and the Comparative result with step 3.
Preferably, step 1 specifically includes: the data of electronic health record include essential information (including admission date, discharge day
Phase, name, age, gender etc.) and diagnosis, saccharification and biochemical analysis data.Essential information and data source hospitalized to have a thorough examination
In different server and database, need to integrate it to obtain the data set that can be used for statisticalling analyze.
Step 2: electronic health record pretreatment includes: data cleansing, data transformation, hough transformation etc..
Data cleansing: after checking that inspection data is matched by the essential information of admission number and patient, discovery data are deposited
It is readable it is not strong, part physical examination analysis data missing is serious, partial data the problems such as there are exceptional values.Noise data is due to disease
Going through is to record form disunity by doctor's hand-kept, and the mode of different doctor's records is different, thus first to data into
The cleanings of data is gone, removal abnormal data, repeated data and existing wrong data.
Data transformation: including several aspect contents, the first feature if necessary is not present, and passes through existing feature calculation
It obtains, secondly the data for being unsatisfactory for standardizing in data is carried out with the rule of processing, such as process range of data normalization etc
Model.
Hough transformation: including many features in initial data, and not all feature requires, and research all pair has
Value, the use of excessive invalid feature can accuracy rate, efficiency to model all have a negative impact.Therefore it reduces without help
Data characteristics is not only helpful to the quality for improving model, and for reducing storage spending, the operational efficiency of model has certain
It helps.
Step 3: the building of diabetes prediction model chooses BP nerve after the characteristics of analyzing electronic health record data
Network exists sensitive to initial weight as basic prediction model according to BP neural network, easily converges on lacking for local minimum
The irregular timing feature of point and diabetes data, improves BP neural network, can preferably be reached afterwards
The effect of prediction.
Step 4: experimental result comparison: while using k nearest neighbor, logistic regression, decision tree, random forests algorithm etc. to place
Diabetes data after reason is predicted, by the comparison of experiment, discovery changes either on training set or on verifying collection
Into BP neural network model all achieve best compliance test result prediction effect of the invention.
Detailed description of the invention
Fig. 1 whole Technology Roadmap proposed by the present invention
The method figure of Fig. 2 Data Integration proposed by the present invention.
Fig. 3 data prediction figure proposed by the present invention.
BP neural network figure Fig. 4 of the invention.
Fig. 5 prediction result comparison diagram of the present invention.
Specific embodiment
As shown in Figure 1, this is general frame of the invention, from the acquisition of data, the pretreatment of data, then prediction model is arrived
The whole process of foundation.Diabetes data is acquired from Hospital Electronic Medical Record first, then according to the quality problems of data for
Diabetes data source has carried out pretreatment operation to obtain target sample data, and uses statistical analysis, the data mining on basis
Means etc. carry out data analysis, the characteristics of to understand data relevant to diabetic complication, and the relevant data characteristics of selection
Deng.And select BP neural network to improve as basic algorithm combined data feature according to the characteristics of diabetes data, and
Diabetic complication risk forecast model is established on the basis of this.
As shown in Figure 1, a kind of diabetes illness risk forecast model construction method based on electronic health record data mining, tool
Body the following steps are included:
Step 1: such as Fig. 2, the electronic health record data come from the export of each server being counted by unique patient's identification number
According to integration, by the essential information of electronic health record, diagnostic message etc. is merged into a complete sample.The step of Data Integration, is such as
Under: the information of diabetes complicated nephrosis and the information of the non-Nephropathy of diabetes are 1. extracted according to diagnostic message for the first time；2. root
According to medical admission number and Diagnostic Time from database saccharification is checked and biochemical analysis in it is nearest apart from admission date
Primary inspection information, which extracts, to be matched.Finally according to the admission number of electronic health record uniquely by the electronic health record number of each section
According to being integrated.
Step 2: such as Fig. 3, electronic health record pretreatment includes: data cleansing, data transformation, hough transformation etc..
By carrying out integration and screening to the data collected, there is also similar spies for the set of data samples by integration
Sign, such as the name of electronic health record are similar with electronic health record ID effect, it is contemplated that the uniqueness of ID and the possibility of name repeat
Property, the present invention selects electronic health record ID to remove the name in electronic health record.Also some is declarer, speaker etc., and chemical examination
Machine variation in information, unrelated can directly delete with establishing for model.To missing be main characteristic information or
The operation that record more than the main characteristic information of missing gives deletion is taken herein for there was only the missing of Individual features value
It is filled up using the method for arest neighbors interpolation and piecewise interpolation.The processing of exceptional value, in initial data, there are mistakes for discovery
Or the data that deviate from desired value, that is, there are some noise datas, will do it deletion or as missing values processing.Pre- place
Reason further includes data transformation and hough transformation, and data transformation is mainly the normalization of data and the discretization of data.Data are returned
One change is normalized using minimax.Discretization using cluster analysis method.Hough transformation is using cluster
Method.
Step 3: the building of diabetes model: such as Fig. 4, this model is mainly using BP neural network as prediction
Model.Because BP neural network is a kind of nonlinear dynamic network, it is non-thread that any one is theoretically approached by network model
Property function, and the inspection inspection data and complication of diabetes are just the case that.And it compared others
Machine learning method, discovery BP neural network is best prediction model.It is very sensitive to initial value for BP neural network, hold
Local extremum easily is fallen into, using the global search of genetic algorithm to initial value and threshold optimization, not otherwise for diabetes data
The timing of rule is added one layer, for saving upper one layer of output, for being fitted medical number in BP neural network hidden layer
According to timing.
Step 4: Comparison of experiment results: such as Fig. 5, the present invention uses the machine learning algorithm K of most study instantly simultaneously
Diabetes data is predicted to treated for neighbour, logistic regression, decision tree, random forests algorithm etc., is tied by experiment
The comparison of fruit and the evaluation criteria of model, discovery is either on training set or verifying collects upper improved BP neural network mould
Type all achieves best compliance test result prediction effect of the invention.
It should be strongly noted that this method is based on the related data that electronic health record uses, not with the people of living body
It carries out or for the purpose for the treatment of, correlation model foundation is all based on data analysis and compares progress, finally obtained model
It is the comparison for serving data, entire technical solution can effectively be implemented, and not medical diagnosis on disease or treatment, this method tool
There is certain technology intensivism, machine processing speed can be greatly improved.
In conclusion the prediction that the present invention has compared diabetes by improved BP neural network, by right
The comparison of conventional machines learning model prediction result highlights the advantage of the algorithm, all from the stability of prediction effect or model
It is more advantageous than general machine learning algorithm, early stage can be carried out to may suffer from diabetes in advance by the prediction of the model
Intervene, reduce the cost cured the disease, also saves resource for hospital.
Claims (8)
1. a kind of diabetes prediction model construction method based on electronic health record data mining, it is characterised in that: the reality of this method
Apply that steps are as follows,
Step 1: the electronic health record data come from the export of each server are carried out to the integration of data by unique patient's identification number, it will
Essential information and diagnostic message of electronic health record etc. are merged into a complete sample；
Step 2: by electronic health record data prediction, obtaining clean available electronic health record data and be stored in database；
Step 3: analyzing resulting electronic health record data cases, several risk factors of diabetes are determined, according to diabetes number
According to the characteristics of the BP neural network of selection is improved, and establish diabetes prediction model on this basis；According to diabetes
The characteristics of data, improves the BP neural network of selection, and establishes diabetes prediction model on this basis, and at
The electronic health record data managed are tested；
Step 4: the training for carrying out model to diabetes data using k nearest neighbor, logistic regression, decision tree, random forests algorithm is pre-
It surveys, and the Comparative result with step 3.
2. a kind of diabetes prediction model construction method based on electronic health record data mining according to claim 1,
Be characterized in that: step 1 specifically includes: the data of electronic health record include essential information include admission date, discharge the date, name,
Age, gender and diagnosis, saccharification and biochemical analysis data；Essential information and data source hospitalized to have a thorough examination are in different services
Device and database need to integrate it to obtain the data set that can be used for statisticalling analyze.
3. a kind of diabetes prediction model construction method based on electronic health record data mining according to claim 1,
Be characterized in that: step 2: electronic health record pretreatment includes: data cleansing, data transformation, hough transformation.
4. a kind of diabetes prediction model construction method based on electronic health record data mining according to claim 3,
Be characterized in that: data cleansing: after checking that inspection data is matched by the essential information of admission number and patient, discovery data are deposited
It is readable it is not strong, part physical examination analysis data missing is serious, partial data the problems such as there are exceptional values；Noise data is due to disease
Going through is to record form disunity by doctor's hand-kept, and the mode of different doctor's records is different, thus first to data into
The cleanings of data is gone, removal abnormal data, repeated data and existing wrong data.
5. a kind of diabetes prediction model construction method based on electronic health record data mining according to claim 3,
Be characterized in that: data transformation: including several aspect contents, the first feature if necessary is not present, and passes through existing feature meter
It obtains, the processing of data normalization is secondly carried out for being unsatisfactory for the data of specification in data.
6. a kind of diabetes prediction model construction method based on electronic health record data mining according to claim 3,
Be characterized in that: hough transformation: comprising many features in initial data, not all feature is required, and is reduced without help
Data characteristics.
7. a kind of diabetes prediction model construction method based on electronic health record data mining according to claim 1,
Be characterized in that: step 3: the building of diabetes prediction model chooses BP nerve after the characteristics of analyzing electronic health record data
Network exists sensitive to initial weight as basic prediction model according to BP neural network, easily converges on lacking for local minimum
The irregular timing feature of point and diabetes data, improves BP neural network, can preferably be reached afterwards
The effect of prediction.
8. a kind of diabetes prediction model construction method based on electronic health record data mining according to claim 1,
It is characterized in that: step 4: experimental result comparison: while k nearest neighbor, logistic regression, decision tree, random forests algorithm are used to place
Diabetes data after reason is predicted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910161929.5A CN109920547A (en) | 2019-03-05 | 2019-03-05 | A kind of diabetes prediction model construction method based on electronic health record data mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910161929.5A CN109920547A (en) | 2019-03-05 | 2019-03-05 | A kind of diabetes prediction model construction method based on electronic health record data mining |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109920547A true CN109920547A (en) | 2019-06-21 |
Family
ID=66963121
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910161929.5A Pending CN109920547A (en) | 2019-03-05 | 2019-03-05 | A kind of diabetes prediction model construction method based on electronic health record data mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109920547A (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111223568A (en) * | 2019-04-18 | 2020-06-02 | 岭南师范学院 | Diabetes early warning model for improving k-means clustering |
| CN111710420A (en) * | 2020-05-15 | 2020-09-25 | 深圳先进技术研究院 | Complication morbidity risk prediction method, system, terminal and storage medium based on electronic medical record big data |
| CN111899866A (en) * | 2020-07-28 | 2020-11-06 | 四川大学华西医院 | Surgical operation complication evaluation system based on deep learning |
| CN111951953A (en) * | 2020-07-20 | 2020-11-17 | 肾泰网健康科技(南京)有限公司 | Method, model and system for screening diabetic kidney disease based on AI technology |
| CN111968748A (en) * | 2020-08-21 | 2020-11-20 | 南通大学 | Modeling method of diabetic complication prediction model |
| CN111986754A (en) * | 2020-08-21 | 2020-11-24 | 南通大学 | Electronic medical record management model construction method based on diabetes |
| CN112164454A (en) * | 2020-10-10 | 2021-01-01 | 联仁健康医疗大数据科技股份有限公司 | Diagnosis prediction method and device and electronic equipment |
| CN113243887A (en) * | 2021-07-16 | 2021-08-13 | 中山大学中山眼科中心 | Intelligent diagnosis and treatment instrument for macular degeneration of old people |
| CN113273959A (en) * | 2021-07-19 | 2021-08-20 | 中山大学中山眼科中心 | Portable diabetic retinopathy diagnosis and treatment instrument |
| WO2021196239A1 (en) * | 2020-04-03 | 2021-10-07 | 清华大学 | Network representation learning algorithm across medical data sources |
| CN113539394A (en) * | 2020-12-31 | 2021-10-22 | 内蒙古卫数数据科技有限公司 | Multi-disease prediction method based on medical inspection data |
| CN113679348A (en) * | 2021-08-26 | 2021-11-23 | 平安国际智慧城市科技股份有限公司 | Blood glucose prediction method, blood glucose prediction device, blood glucose prediction apparatus, and storage medium |
| CN116486921A (en) * | 2023-03-31 | 2023-07-25 | 北京大学深圳医院（北京大学深圳临床医学院） | Breast cancer risk prediction method based on electronic medical record big data |
| CN116779190A (en) * | 2023-06-25 | 2023-09-19 | 急尼优医药科技(上海)有限公司 | Medical platform user follow-up management system and method based on Internet of things |
| JP7404581B1 (en) | 2022-08-16 | 2023-12-25 | 之江実験室 | Chronic nephropathy subtype mining system based on self-supervised graph clustering |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014149827A1 (en) * | 2013-03-15 | 2014-09-25 | REMTCS Inc. | Artificial neural network interface and methods of training the same for various use cases |
| WO2016168090A1 (en) * | 2015-04-14 | 2016-10-20 | Nueon, Inc. | Method and apparatus for determining markers of health by analysis of blood |
| CN106503802A (en) * | 2016-10-20 | 2017-03-15 | 上海电机学院 | A kind of method of utilization genetic algorithm optimization BP neural network system |
| CN109346177A (en) * | 2018-09-10 | 2019-02-15 | 昆明理工大学 | A kind of disease symptoms quantitative forecasting technique based on improvement GA-BP algorithm |
-
2019
- 2019-03-05 CN CN201910161929.5A patent/CN109920547A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014149827A1 (en) * | 2013-03-15 | 2014-09-25 | REMTCS Inc. | Artificial neural network interface and methods of training the same for various use cases |
| WO2016168090A1 (en) * | 2015-04-14 | 2016-10-20 | Nueon, Inc. | Method and apparatus for determining markers of health by analysis of blood |
| CN106503802A (en) * | 2016-10-20 | 2017-03-15 | 上海电机学院 | A kind of method of utilization genetic algorithm optimization BP neural network system |
| CN109346177A (en) * | 2018-09-10 | 2019-02-15 | 昆明理工大学 | A kind of disease symptoms quantitative forecasting technique based on improvement GA-BP algorithm |
Non-Patent Citations (2)
| Title |
|---|
| 李瀚阳等: "基于GA-BP神经网络的手足口病预测模型", 《滨州学院学报》 * |
| 王俊杰等: "BP神经网络在疾病预测中的应用", 《数理医药学杂志》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111223568B (en) * | 2019-04-18 | 2024-03-26 | 岭南师范学院 | Diabetes early warning model for improving k-means clustering |
| CN111223568A (en) * | 2019-04-18 | 2020-06-02 | 岭南师范学院 | Diabetes early warning model for improving k-means clustering |
| WO2021196239A1 (en) * | 2020-04-03 | 2021-10-07 | 清华大学 | Network representation learning algorithm across medical data sources |
| CN111710420A (en) * | 2020-05-15 | 2020-09-25 | 深圳先进技术研究院 | Complication morbidity risk prediction method, system, terminal and storage medium based on electronic medical record big data |
| CN111710420B (en) * | 2020-05-15 | 2024-03-19 | 深圳先进技术研究院 | Complication onset risk prediction method, system, terminal and storage medium based on electronic medical record big data |
| WO2021227511A1 (en) * | 2020-05-15 | 2021-11-18 | 深圳先进技术研究院 | Complication onset risk prediction method and system based on electronic medical record big data, and terminal and storage medium |
| CN111951953A (en) * | 2020-07-20 | 2020-11-17 | 肾泰网健康科技(南京)有限公司 | Method, model and system for screening diabetic kidney disease based on AI technology |
| CN111899866B (en) * | 2020-07-28 | 2022-04-22 | 四川大学华西医院 | Surgical operation complication evaluation system based on deep learning |
| CN111899866A (en) * | 2020-07-28 | 2020-11-06 | 四川大学华西医院 | Surgical operation complication evaluation system based on deep learning |
| CN111968748A (en) * | 2020-08-21 | 2020-11-20 | 南通大学 | Modeling method of diabetic complication prediction model |
| CN111986754A (en) * | 2020-08-21 | 2020-11-24 | 南通大学 | Electronic medical record management model construction method based on diabetes |
| CN112164454A (en) * | 2020-10-10 | 2021-01-01 | 联仁健康医疗大数据科技股份有限公司 | Diagnosis prediction method and device and electronic equipment |
| CN113539394A (en) * | 2020-12-31 | 2021-10-22 | 内蒙古卫数数据科技有限公司 | Multi-disease prediction method based on medical inspection data |
| CN113243887A (en) * | 2021-07-16 | 2021-08-13 | 中山大学中山眼科中心 | Intelligent diagnosis and treatment instrument for macular degeneration of old people |
| CN113273959A (en) * | 2021-07-19 | 2021-08-20 | 中山大学中山眼科中心 | Portable diabetic retinopathy diagnosis and treatment instrument |
| CN113679348A (en) * | 2021-08-26 | 2021-11-23 | 平安国际智慧城市科技股份有限公司 | Blood glucose prediction method, blood glucose prediction device, blood glucose prediction apparatus, and storage medium |
| CN113679348B (en) * | 2021-08-26 | 2024-02-06 | 深圳平安智慧医健科技有限公司 | Blood glucose prediction method, blood glucose prediction device, apparatus, and storage medium |
| JP7404581B1 (en) | 2022-08-16 | 2023-12-25 | 之江実験室 | Chronic nephropathy subtype mining system based on self-supervised graph clustering |
| CN116486921A (en) * | 2023-03-31 | 2023-07-25 | 北京大学深圳医院（北京大学深圳临床医学院） | Breast cancer risk prediction method based on electronic medical record big data |
| CN116779190A (en) * | 2023-06-25 | 2023-09-19 | 急尼优医药科技(上海)有限公司 | Medical platform user follow-up management system and method based on Internet of things |
| CN116779190B (en) * | 2023-06-25 | 2024-02-13 | 急尼优医药科技(上海)有限公司 | Medical platform user follow-up management system and method based on Internet of things |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109920547A (en) | A kind of diabetes prediction model construction method based on electronic health record data mining | |
| Yuan et al. | The development an artificial intelligence algorithm for early sepsis diagnosis in the intensive care unit | |
| Christo et al. | Feature selection and instance selection from clinical datasets using co-operative co-evolution and classification using random forest | |
| ȚĂRANU | Data mining in healthcare: decision making and precision | |
| Sk et al. | Coronary Heart Disease Prediction and Classification using Hybrid Machine Learning Algorithms | |
| CN107785057B (en) | Medical data processing method, device, storage medium and computer equipment | |
| Rahman et al. | Using and comparing different decision tree classification techniques for mining ICDDR, B Hospital Surveillance data | |
| CN112669967B (en) | Active health medical decision-making assisting method and equipment | |
| CN106599553A (en) | Disease early-warning method and device | |
| Pal et al. | Deep learning techniques for prediction and diagnosis of diabetes mellitus | |
| CN107145715B (en) | Clinical medicine intelligence discriminating gear based on electing algorithm | |
| Zhang et al. | Medical diagnosis data mining based on improved Apriori algorithm | |
| Khajehei et al. | Data mining and medical research studies | |
| CN110046757B (en) | Outpatient clinic volume prediction system and prediction method based on LightGBM algorithm | |
| CN102405473A (en) | A point-of-care enactive medical system and method | |
| CN113838577B (en) | Convenient layered old people MODS early death risk assessment model, device and establishment method | |
| CN112967803A (en) | Early mortality prediction method and system for emergency patients based on integrated model | |
| CN109243604A (en) | A kind of construction method and building system of the Kawasaki disease risk evaluation model based on neural network algorithm | |
| Zhang et al. | Identifying diabetic macular edema and other retinal diseases by optical coherence tomography image and multiscale deep learning | |
| Chou et al. | Extracting drug utilization knowledge using self-organizing map and rough set theory | |
| Mirzajani | Prediction and diagnosis of diabetes by using data mining techniques | |
| JP7365747B1 (en) | Disease treatment process abnormality identification system based on hierarchical neural network | |
| CN109192306A (en) | A kind of judgment means of diabetes, equipment and computer readable storage medium | |
| CN108538390A (en) | A kind of increment type processing method towards medical data | |
| Gopalakrishnan et al. | A Novel Deep Learning-Based Heart Disease Prediction System Using Convolutional Neural Networks (CNN) Algorithm |
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: 20190621 |
|
| RJ01 | Rejection of invention patent application after publication |