CN109830303A - Clinical data mining analysis and aid decision-making method based on internet integration medical platform - Google Patents
Clinical data mining analysis and aid decision-making method based on internet integration medical platform Download PDFInfo
- Publication number
- CN109830303A CN109830303A CN201910101985.XA CN201910101985A CN109830303A CN 109830303 A CN109830303 A CN 109830303A CN 201910101985 A CN201910101985 A CN 201910101985A CN 109830303 A CN109830303 A CN 109830303A
- Authority
- CN
- China
- Prior art keywords
- data
- analysis
- module
- clinical
- algorithm
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 140
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000007418 data mining Methods 0.000 title claims abstract description 43
- 230000010354 integration Effects 0.000 title claims abstract description 27
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 77
- 238000007405 data analysis Methods 0.000 claims abstract description 42
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims abstract description 34
- 238000012550 audit Methods 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 28
- 201000010099 disease Diseases 0.000 claims abstract description 26
- 238000013135 deep learning Methods 0.000 claims abstract description 12
- 238000011156 evaluation Methods 0.000 claims description 25
- 238000012545 processing Methods 0.000 claims description 25
- 238000000611 regression analysis Methods 0.000 claims description 22
- 238000013528 artificial neural network Methods 0.000 claims description 20
- 238000012360 testing method Methods 0.000 claims description 18
- 238000006243 chemical reaction Methods 0.000 claims description 15
- 238000013527 convolutional neural network Methods 0.000 claims description 13
- 238000010801 machine learning Methods 0.000 claims description 13
- 238000007637 random forest analysis Methods 0.000 claims description 13
- 238000003066 decision tree Methods 0.000 claims description 12
- 238000005065 mining Methods 0.000 claims description 11
- 239000000203 mixture Substances 0.000 claims description 11
- 230000000306 recurrent effect Effects 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 11
- 238000005516 engineering process Methods 0.000 claims description 9
- 238000012417 linear regression Methods 0.000 claims description 9
- 238000010219 correlation analysis Methods 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 8
- 238000012098 association analyses Methods 0.000 claims description 7
- 238000012706 support-vector machine Methods 0.000 claims description 7
- 239000000284 extract Substances 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 5
- 230000007787 long-term memory Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 230000001419 dependent effect Effects 0.000 claims description 4
- 238000003064 k means clustering Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 230000010365 information processing Effects 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000000465 moulding Methods 0.000 claims description 3
- 238000011524 similarity measure Methods 0.000 claims description 3
- 241001074085 Scophthalmus aquosus Species 0.000 claims description 2
- 238000004321 preservation Methods 0.000 claims description 2
- 230000001537 neural effect Effects 0.000 claims 1
- 206010012601 diabetes mellitus Diseases 0.000 abstract description 29
- 208000026310 Breast neoplasm Diseases 0.000 abstract description 19
- 206010020850 Hyperthyroidism Diseases 0.000 abstract description 19
- 208000024770 Thyroid neoplasm Diseases 0.000 abstract description 19
- 238000011160 research Methods 0.000 abstract description 17
- 208000009453 Thyroid Nodule Diseases 0.000 abstract description 16
- 238000013480 data collection Methods 0.000 abstract description 7
- 238000003759 clinical diagnosis Methods 0.000 abstract description 3
- 238000003745 diagnosis Methods 0.000 description 24
- 239000000523 sample Substances 0.000 description 20
- 210000001685 thyroid gland Anatomy 0.000 description 18
- 230000008569 process Effects 0.000 description 17
- 238000007689 inspection Methods 0.000 description 14
- 230000036541 health Effects 0.000 description 11
- 206010028980 Neoplasm Diseases 0.000 description 10
- 238000009412 basement excavation Methods 0.000 description 9
- 238000007726 management method Methods 0.000 description 9
- 238000012731 temporal analysis Methods 0.000 description 9
- 208000021510 thyroid gland disease Diseases 0.000 description 9
- 238000000700 time series analysis Methods 0.000 description 9
- 208000024799 Thyroid disease Diseases 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 241001269238 Data Species 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 208000029078 coronary artery disease Diseases 0.000 description 6
- 239000003814 drug Substances 0.000 description 6
- 238000002604 ultrasonography Methods 0.000 description 6
- 238000013473 artificial intelligence Methods 0.000 description 5
- 210000000481 breast Anatomy 0.000 description 5
- 230000008859 change Effects 0.000 description 5
- 210000005075 mammary gland Anatomy 0.000 description 5
- 230000007170 pathology Effects 0.000 description 5
- 238000007781 pre-processing Methods 0.000 description 5
- 238000002360 preparation method Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 229940079593 drug Drugs 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 235000019580 granularity Nutrition 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 230000000505 pernicious effect Effects 0.000 description 4
- 230000009467 reduction Effects 0.000 description 4
- 208000013076 thyroid tumor Diseases 0.000 description 4
- WVCHIGAIXREVNS-UHFFFAOYSA-N 2-hydroxy-1,4-naphthoquinone Chemical compound C1=CC=C2C(O)=CC(=O)C(=O)C2=C1 WVCHIGAIXREVNS-UHFFFAOYSA-N 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000004140 cleaning Methods 0.000 description 3
- 238000013075 data extraction Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 230000002526 effect on cardiovascular system Effects 0.000 description 3
- 230000002124 endocrine Effects 0.000 description 3
- 230000008676 import Effects 0.000 description 3
- 230000001575 pathological effect Effects 0.000 description 3
- 238000000513 principal component analysis Methods 0.000 description 3
- 238000003672 processing method Methods 0.000 description 3
- 238000012552 review Methods 0.000 description 3
- 210000002966 serum Anatomy 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 238000006467 substitution reaction Methods 0.000 description 3
- FDORQEIHOKEJNX-UHFFFAOYSA-N 2-[[3-(4-fluorophenyl)-3-(4-phenylphenoxy)propyl]-methylamino]acetic acid Chemical compound C=1C=C(F)C=CC=1C(CCN(C)CC(O)=O)OC(C=C1)=CC=C1C1=CC=CC=C1 FDORQEIHOKEJNX-UHFFFAOYSA-N 0.000 description 2
- XUIIKFGFIJCVMT-GFCCVEGCSA-N D-thyroxine Chemical compound IC1=CC(C[C@@H](N)C(O)=O)=CC(I)=C1OC1=CC(I)=C(O)C(I)=C1 XUIIKFGFIJCVMT-GFCCVEGCSA-N 0.000 description 2
- AUYYCJSJGJYCDS-LBPRGKRZSA-N Thyrolar Chemical compound IC1=CC(C[C@H](N)C(O)=O)=CC(I)=C1OC1=CC=C(O)C(I)=C1 AUYYCJSJGJYCDS-LBPRGKRZSA-N 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 210000004027 cell Anatomy 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000013501 data transformation Methods 0.000 description 2
- 238000000586 desensitisation Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000013508 migration Methods 0.000 description 2
- 230000005012 migration Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 229940034208 thyroxine Drugs 0.000 description 2
- XUIIKFGFIJCVMT-UHFFFAOYSA-N thyroxine-binding globulin Natural products IC1=CC(CC([NH3+])C([O-])=O)=CC(I)=C1OC1=CC(I)=C(O)C(I)=C1 XUIIKFGFIJCVMT-UHFFFAOYSA-N 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 241000894006 Bacteria Species 0.000 description 1
- 206010006187 Breast cancer Diseases 0.000 description 1
- 208000002249 Diabetes Complications Diseases 0.000 description 1
- 206010012655 Diabetic complications Diseases 0.000 description 1
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 206010018473 Glycosuria Diseases 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000036772 blood pressure Effects 0.000 description 1
- 230000036760 body temperature Effects 0.000 description 1
- 238000010224 classification analysis Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 239000006071 cream Substances 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 229940088597 hormone Drugs 0.000 description 1
- 239000005556 hormone Substances 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000003455 independent Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000474 nursing effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000010827 pathological analysis Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 238000005086 pumping Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000001646 thyrotropic effect Effects 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 229940035722 triiodothyronine Drugs 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 238000007794 visualization technique Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Abstract
Present invention discloses a kind of clinical data mining analysis and aid decision-making method based on internet integration medical platform, it is related to internet medical platform technical field, including data mining analysis and aid decision, data mining analysis includes multidimensional analysis algoritic module, data mining algorithm module, deep learning algoritic module；Aided remote decision is made of four parts such as the prediction module based on index parameter, the prediction module based on audit report text, model training module and structurized modules.The present invention selects the research object of hyperthyroidism, diabetes, thyroid nodule, several diseases of tumor of breast as data collection and analysis, unified platform acquisition is relied on to integrate clinical medical data, it realizes the data mining analysis towards the diseases clinical data such as hyperthyroidism, diabetes, thyroid nodule, tumor of breast and aid decision service, system is provided and is supported for clinician's clinical diagnosis and scientific research personnel's disease research.
Description
Technical field
The present invention relates to internet medical platform technical fields, more specifically refer to a kind of flat based on internet integration medical treatment
The clinical data mining analysis and aid decision-making method of platform.
Background technique
Big data penetrates into each industry and department, depth is answered as a kind of important resource to some extent
With the business activities for not only facilitating constituent parts, it is also beneficial to push the development of national economy." internet+" be various countries' industry and
The achievement and mark of information-based depth integration, and further promote the important handgrip of information consumption.So-called " internet+" is exactly
" internet+each traditional industries ", but this is not both simple addition, but utilize Information and Communication Technology and internet
Platform allows internet and traditional industries to carry out depth integration, creates new developing ecology.Future Internet also can as electricity,
As a kind of productivity tool, being substantially improved for efficiency is brought to each industry.Push mobile Internet, cloud computing, big data,
Internet of Things etc. promotes e-commerce, industry internet and the development of internet financial health, such as in conjunction with modern manufacturing industry " tradition
Fairground+internet, traditional general merchandise sales field+internet, traditional bank+internet, traditional matchmaker+internet have, and tradition is handed over
Logical+internet." internet+" it is financial, mutual to form such as internet medical treatment, internet just in overall application to the tertiary industry
The new industry situations such as networking traffic, Internet education.
Medical industry is the important component of national economy and social development, and under the new situation, medical information is built
Fast development have benefited from the application of the IT emerging technology such as big data, cloud computing and Internet of Things, caused the big of medical data
Explosion, promotes the formation of medical big data.Traditional medical is not overturned in internet, but is with internet, mobile Internet
Means are examined to divide, Extension of service radius, to solve the problems, such as medical resource insufficient supply and be unevenly distributed weighing apparatus.Internet medical treatment
The communication capability between patient, medical service organ and doctor is improved, traditional Site Service mode is broken through, alleviates medical treatment
The status of scarcity of resources.But do not got through between hospital and with the information sharing outside institute in institute, problem of detached island of information is significant,
A degree of restriction is brought for the effective use of medical data, the user health data being well worth doing originally become nothing and use force it
Ground.
Summary of the invention
(1) the technical issues of solving
It is an object of the present invention to provide a kind of clinical data mining analysis based on internet integration medical platform and auxiliary
Decision-making technique selects the research of hyperthyroidism, diabetes, tumor of breast and several diseases of thyroid tumors as data collection and analysis
Object relies on unified platform acquisition to integrate clinical medical data, realizes towards hyperthyroidism, diabetes, tumor of breast and thyroid gland
The data mining analysis of the diseases clinical data such as tumour and aid decision service are clinician's clinical diagnosis and scientific research personnel's disease
Disease research offer system is supported.
(2) technical solution
Clinical data mining analysis and aid decision-making method based on internet integration medical platform, including data mining
Analysis and aid decision, data mining analysis include multidimensional analysis algoritic module, data mining algorithm module, deep learning algorithm
Module；Multidimensional analysis algoritic module chooses several numbers firstly the need of cube, cube is established from data warehouse
According to subset, then organize and be aggregated into the multidimensional structure as defined in multiple dimensions and metric；Data mining algorithm module provides
Uniform registration including machine learning algorithms such as classification, cluster, correlation rule, regression analyses is used for using managing with nullifying
For the mining analysis of specific set of data, clinical depth analysis, early warning and prediction are realized；Deep learning prediction algorithm module collection
At the recurrent neural networks model of convolutional neural networks (CNN), Recognition with Recurrent Neural Network (RNN) and shot and long term memory unit
(LSTM) scheduling algorithm；Aided remote decision by the prediction module based on index parameter, the prediction module based on audit report text,
Model training module and structurized module composition.
An embodiment according to the present invention, the multidimensional analysis algoritic module are vertical to the data organized with multi-dimensional form
Cube carries out volume, lower brill, slice, stripping and slicing, a variety of analyses of rotation operate, so as to profile data, enable analyst, policymaker from
Data in multiple angles, multiple sides observation database, so that understanding in depth includes information and intension in data.
An embodiment according to the present invention, the multidimensional analysis algoritic module include dimension and metric, and the dimension is
User observes the angle of data, and what it is comprising dimensional information is dimension table, detailed value or factual data comprising preservation metric
It is true table, dimension table includes the characteristic of the true record in description fact table；The metric is with practical significance
Numerical value, metric is stored in true table.
An embodiment according to the present invention, it is described to carry out model training using data mining algorithm, it is broadly divided into problem
Analysis, data prediction, data modeling, outcome evaluation, knowledge apply five steps:
(1) case study: by the productive discussions with doctor, expert, the target of explicit data analysis defines data analysis
Demand, setting data analysis expected results；
(2) data prediction: determine data analysis data source, construct reasonable data warehouse for data inquiry with
Analysis, using sqlserver, mysql database, java, python high-level language, R, spss statistical means carry out data
Necessary pretreatment work；
(3) using processed data as input feature vector, suitable machine learning model, Huo Zhe data modeling: are selected
It is suitably modified on the basis of existing machine learning model, using objective attribute target attribute as output feature, during model training
Parameter adjustment and model optimization is repeated, obtains the optimal training pattern of effect；
(4) outcome evaluation: after the completion of model construction, it would be desirable to carry out the assessment of science, to the reliability of model with true
The convincingness for protecting experimental result, specifically, using accuracy rate, can be recalled by the methods of ROC curve and confusion matrix
The indexs such as rate, F1 value measure the quality of model；
(5) it knowledge application: goes to predict unknown things using molding model, or seeks unknown things and known things
Between connection, so that people be helped preferably to recognize unknown things.
An embodiment according to the present invention, be integrated in the data mining algorithm module random forest, support vector machines,
Neural network, regression analysis, correlation analysis, Apriori association analysis, K-means clustering algorithm；
(1) random forests algorithm is described in detail below: firstly, being concentrated with the pumping put back to from sample using bagging technology
The subset of same size is taken, repeats K times, that is, generates K sample set, it is then right using each self-generating of these sample sets one
The decision tree answered generates optimal decision tree to each sample set, finally carries out the classification prediction result of these decision trees
Ballot, classification prediction result of the several most results of getting tickets as Random Forest model；
(2) algorithm of support vector machine finds an optimal segmentation hyperplane, so that the plane is guaranteeing to maximize classification
Under the premise of accuracy rate, the distance of both sides sample to hyperplane is maximum；
(3) neural network algorithm can be divided into several component parts in structure, be input layer, hidden layer and output respectively
Layer, wherein input layer receives the initial data of extraneous input, and passes it to hidden layer；Hidden layer may include one layer or
Multilayer is responsible for internal information processing and conversion, and the information after conversion is passed to output layer；Information is in output layer by most
After processing and conversion afterwards, final result is outwardly exported；
(4) algorithm with regress analysis method is divided into simple regression analysis and polynary according to the quantity of variable involved in analyzing
Regression analysis；According to the quantity of independent variable, linear regression analysis can be divided into simple regression analysis and multiple regression point
Analysis；According to the relationship type between independent variable and dependent variable, linear regression analysis and nonlinear regression analysis can be divided into；
(5) it whether there is certain dependence between correlation analysis algorithm data object, and have dependence for specific
The data object of relationship inquires into its related direction and degree of correlation；
(6) Apriori association analysis algorithm, the first step retrieve all frequent episodes in transaction database by iteration
Collection；Second step constructs the rule for meeting user's min confidence using frequent item set；
(7) for K-means algorithm using manhatton distance or Euclidean distance as similarity measure, it is to ask corresponding a certain first
Beginning cluster centre vector v most has classification, so that evaluation index J is minimum.
An embodiment according to the present invention, the prediction module based on index parameter: according to patient's outpatient service serial number or
Person's medical insurance card number information inquires the test rating and audit report text of the related disease of the patient, for structuring index
Data can directly input；For non-structured audit report, obtaining model using the progress structuring of structuring submodule can know
Other data format；
The prediction module based on audit report text: according to the information such as patient's outpatient service serial number or medical insurance card number,
Inquire the audit report text of the patient；For non-structured audit report text, directly using deep learning algorithm into
Row prediction；
The model training module: by the relevant clinical audit report of multiple database medical information systems, test rating
Data merge processing, are integrated into unified tables of data, carry out model training；
The structurized module: it realizes that the structuring to ultrasonic report text data is handled, extracts the ultrasound of various samples
The index value of feature and each index, and form the description template by each sample.
(3) beneficial effect
Using technical solution of the present invention, clinical data mining analysis based on internet integration medical platform with it is auxiliary
Decision-making technique is helped, unstructured clinical document is input to clinical document structuring processing engine, passes through clinical medicine corpus, rule
Then, the means such as full-text search and machine learning are handled, and are obtained structural data and are output to distributed storage engine, pass through artificial intelligence
Energy algorithm is handled, and for Platform Analysis, is shown；Text data non-structured in clinical data is carried out structure by the present invention
Change processing, stores into distributed Hadoop cluster, realizes Distributed Storage mode and distributed computing processing, and will be
Programming in software application, which is realized, to be transformed and is adapted to for distributed nature.
Detailed description of the invention
In the present invention, identical appended drawing reference always shows identical feature, in which:
Fig. 1 is integrated medical platform general frame figure Internet-based.
Fig. 2 is hyperthyroidism clinical medical data analysis system level architecture diagram.
Fig. 3 is that diabetes clinical data analysis excavates general frame figure.
Fig. 4 is that thyroid nodule clinical data analysis excavates integrated stand composition.
Fig. 5 is the thyroid disease classification method flow chart based on random forest.
Specific embodiment
Technical solution of the present invention is further illustrated with reference to the accompanying drawings and examples.
Integration medical platform Internet-based combines medical big data and artificial intelligence technology, realizes based on " interconnection
The integrated big data medical services platform of net+medical treatment ", for all participation health cares, movable personal and mechanism provides data
The medical services of the online health care new model such as shared, business operation and cooperation with service, optimization information communication, advantageously promote
Doctors and patients' information mutual communication facilitates service and management that hospital improves itself.Platform general frame is as shown in Figure 1, in platform under
It is supreme to be respectively as follows: platform data basal layer, data analysis layer, medical information resource layer, data depth application layer and client layer etc.
Five levels.Integration medical services platform Internet-based, including back-stage management end, doctor terminal and the big portion of patient end three
Point.
Integrated medical services backstage management of platform end:
Back-stage management provides hospital HIS, the data exchanges such as PACS, LIS, RIS integration, medical information system medical data
The functions such as backup.Mainly by data pick-up integration, medical data backup storage, special population database and anonymous public medical record number
The composition such as inquiry according to library.
(1) data pick-up is integrated: completing the mistake of extraction, conversion and the load of the system datas such as HIS, RIS, LIS, PACS
The clinical data that different clinic information systems generate is carried out unified integration and summarized, realized and suffer from different clinical information by journey
The unification of person's mark and the unification of patient clinical information, make clinical data can unify storage.
A) HIS data extraction module, which is realized, registers, goes to a doctor, examines from HIS Emergency call and HIS system increment extraction of being hospitalized
Break, doctor's advice, be admitted to hospital, the clinical datas such as expense；
B) RIS data extraction module realizes from RIS system increment synchronization audit report, position detail etc. and checks data；
C) LIS data extraction module is realized from LIS system increment synchronization survey report, test rating, bacterium and susceptibility
Etc. inspection datas；
D) PACS module realizes the access from image documentation equipment such as DR, CT etc. the data for following DICOM3.0 consensus standard.
E) ETL subsystem is completed to operate desensitization, cleaning and conversion of clinical data etc..
Data desensitization: desensitizing for patient individual's sensitive data, and patient identity card number, medical card number, patient are personal
Name etc. carries out specially treated, removes sensitive composition.
Data cleansing: incomplete data are abandoned；The data wrong for format, such as date of birth, pass through
Other related datas are repaired, and can not repair, data are marked；
Data conversion: to the enumerated value for using numerical value or character to save in the system of source, the text of corresponding meaning is converted to.
(2) medical data backup storage: medical data backup center is the basis of clinical big data storage, for clinical big number
Initial data source is provided according to processing, analysis.System is using distributed Hadoop cloud storage architecture, and for different medical, mechanism is provided
The distributed storage ability of linear expansion, realize data storage filing, management and shared and all types medical institutions it
Between information intercommunication, shared, achieve the purpose that the diversification storage and access of cloud computing platform.Medical data backup center is by curing
The modules compositions such as treatment data bulk migration, medical data increment import, medical data is checked.
A) medical data bulk migration: use hadoop distributed structure/architecture, realize medical information system medical data by
The monolithic backup that time carries out.
B) medical data increment imports: in the incremental mode of time series, the medical treatment imported in medical information system increases
Measure data.
C) medical data is checked: being realized to kinds of Diseases, Gender, age bracket, department, audit report type and inspection
Time etc. imports medical data and is inquired.
(3) it special population database: according to the patient clinical data of medical information system, establishes towards hyperthyroidism, glycosuria
The special population database of the diseases such as disease, thyroid nodule, tumor of breast and thyroid tumors, can be to kinds of Diseases, patient
Gender, age, inspection doctor, Index for examination and review time etc. inquire.
(4) state of an illness case and the doctor of the patients such as diabetes, thyroid disease anonymous public clinical record data base: can be checked
Diagnosis and treatment suggestion, see a doctor to the patient of the similar state of an illness and reference be provided.In view of privacy, number is established using anonymous form for patient
According to library.Kinds of Diseases, illness description content, doctor can be suggested in detail, check doctor, enquirement and time for replying etc. to look into
It askes.
(5) model library: in order to which the model constructed using intelligent algorithm carries out classification forecast analysis, mould to medical diagnosis on disease
The management of artificial intelligence model is mainly realized in type library, including importing, model training and model such as check at the functions.
(6) system administration: unified platform is mainly directed towards information centre, medical institutions administrative staff, doctor and patient etc. no
With role, need scientifically to manage these users, lead to user management and role rights management, to it is various operation with
Data access authority carries out stringent authorization and control.
Integrated medical services platform doctor terminal:
Integrated medical services platform doctor terminal is mainly that the medical personnel of medical institutions and researcher provide medicine
Research and medical diagnosis aid decision provide platform, establish doctors and patients' channel of communication, check the medical advice of patient and for diagnosis
Evaluation.Mainly by special population analysis, aided remote decision, patient advisory checks, evaluation of patient is checked etc. forms.
(1) special population is analyzed: to the clinical data for suffering from the special populations such as hyperthyroidism and diabetes in hospital information system
Analysis mining is carried out disease research is provided and is provided and is for clinician and scientific research personnel to obtain occurrence regularity and inherent mechanism
System is supported.
A) hyperthyroidism clinical data analysis excavates: hyperthyroidism clinical data includes the medical note of the Basic Information Table of patient, patient
The clinical datas tables such as table, the medicining condition table of patient, the index test table of patient and the diagnosis situation table of patient are recorded, number is recorded
Total amount about 2,000,000.Realize and data mining analysis carried out to the clinical data of hyperthyroidism disease, mainly from the essential information of patient,
The themes such as test rating data information, doctor's advice medicining condition, complication situation, recurrence carry out.
B) diabetes clinical data analysis excavates: Basic Information Table of the diabetes clinical data comprising patient, patient are just
The clinical datas tables such as record sheet, the medicining condition table of patient, the index test table of patient and the diagnosis situation table of patient are examined, are remembered
Record number total amount about 1,000,000.It realizes and data mining analysis is carried out to the clinical data of hyperthyroidism disease, mainly from the basic letter of patient
The themes such as breath, test rating data information, doctor's advice medicining condition, diagnosis situation carry out.
(2) aided remote decision: selection endocrine subject, the thyroid gland of cardiovascular subject and tumour subject, coronary heart disease and
Research object of several diseases such as tumour as data collection and analysis relies on unified platform acquisition to integrate clinical treatment number
According to realizing the medical diagnosis aid decision-making system towards thyroid nodule, coronary heart disease and tumor of breast etc., face for clinician
Bed diagnosis and scientific research personnel's disease research provide system and support.Mainly by based on index parameter prediction module, based on check report
Accuse four parts such as prediction module, model training module and the structurized module of text composition.
A) based on the prediction module of index parameter: according to the information such as patient's outpatient service serial number or medical insurance card number, Ke Yicha
Ask the test rating and audit report text of the related disease of the patient.Structuring achievement data can be directly inputted；It is right
In non-structured audit report, structuring is carried out using structuring submodule and obtains the data format that model can identify.
B the prediction module) based on audit report text:, can according to the information such as patient's outpatient service serial number or medical insurance card number
To inquire the audit report text of the patient.For non-structured audit report text, deep learning algorithm is directly utilized
It is predicted.
C) model training module: belonging to the basic module of system, invisible to user.By the thyroid gland knot of multiple databases
The data such as the relevant clinical audit reports of the medical information systems such as section, coronary heart disease, mammary gland, test rating merge processing, collect
At into unified tables of data, model training is carried out.
D) structurized module: realize that the structuring to ultrasonic report text data is handled, the ultrasound for extracting various samples is special
Sign includes the index value of Tumor size, boundary, echo distribution, echo intensity etc. and each index, and forms retouching by each sample
State template.Based on the template, the processing of the structuring to ultrasonic content of text is realized.
(3) patient advisory checks: it realizes doctor and conditions of patients diagnosis consulting content is checked, it can be according to disease kind
Class, illness description content, review time etc. screening are checked.
(4) evaluation of patient is checked: it realizes doctor and evaluation of patient content is checked, it can be according to physician names, patient
Name, evaluation content, evaluation time etc. screening are checked.
Integrated medical services platform patient end:
Integrated medical services platform patient end is the interface that patient logs in platform, and predominantly patient provides remotely cures on line
Service is treated, evaluates service, the inquiry of Patients ' Electronic health account etc. after being mainly included in line consulting interrogation, medical treatment.Patient can lead to
Online interrogation is crossed, the state of an illness tentative diagnosis result that artificial intelligence technology provides is obtained；, clothes horizontal by on-line evaluation doctor medical skill
Attitude of being engaged in etc.；Diagnosis, inspection, inspection and image, doctor's advice, medical history, pathology and expense etc. are checked by Patients ' Electronic health account
Data.It mainly include three modules: evaluation service, Patients ' Electronic health account after patient advisory's service, medical treatment.
(1) patient advisory services: realizing and provides online interrogation service for patient.Patient provides original state of an illness symptom and retouches
It states, data, the system such as image check text report, test rating value obtain model energy using OCR identification facility, structured techniques
The data format of identification examines unknown sample using the model that intelligent algorithm constructs by test rating signature analysis
It is disconnected to carry out classification prediction, the state of an illness result for predicting the patient is finally showed into patient, including thyroid nodule type, thyroid gland
Good pernicious, Breast Tumors of type of surgery, thyroid tumors etc. achieve the purpose that instruct patient's medical treatment and health care.System
Be integrated with including convolutional neural networks (CNN), Recognition with Recurrent Neural Network (RNN), shot and long term memory unit recurrent neural network mould
The intelligent algorithms such as type (LSTM), random forest, support vector machines, neural network, decision tree and K-means, construct
Thyroid nodule and Breast Tumor Patients disease auxiliary diagnosis prediction model.
(2) service is evaluated after medical treatment: realizing rear evaluation of the patient to doctor's diagnosis and treatment process.Patient on the line of doctor to commenting
Valence is a kind of effective doctor patient communication channel, is improved service quality for medical institutions and doctor, and gradually alleviating conflict between doctors and patients is
There is great help.Doctor can be according to the evaluation and demand of patient come improvement, and medical institutions can be according to patient to doctor
Overall evaluation situation give rewards and punishments appropriate.But the review number of single doctor may just have hundreds and thousands of in practice,
Doctor's quantity of one medical institutions has several hundred or even thousands of, it will generates the evaluation of patient text information of magnanimity, manually
Method needs to expend a large amount of energy to handle and analyze these information.System realizes the medical care evaluation body based on artificial intelligence
System carries out emotional semantic analysis to evaluation of patient by machine, identifies front and unfavorable ratings automatically, count proportion.
Doctor can quickly filter out unfavorable ratings, make improvement according to content；Medical institutions can be by department, doctor etc. to magnanimity
Overall evaluation situation statistical analysis is carried out in evaluation information.
(3) Patients ' Electronic health account:
The clinical data for relying on Data Integration module to generate the different clinic information system such as HIS, PACS, LIS, RIS into
Row and summarizes at unified integration, establish include patient essential information, diagnosis, inspection, inspection and image, doctor's advice, medical history, pathology
With the personal electric health account unified view view of the data such as expense, it can be convenient patient and have access at any time, be diagnosis and treatment and scientific research
Application is provided using clinical big data to support.
A) patient basis's dimension: name, gender, date of birth, passport NO., the contact method of main display patient
Etc. essential informations；
B it) diagnoses dimension: showing all previous diagnosis records of patient etc.；
C it) examines dimension: showing all previous inspection record of patient in a tabular form；
D ultrasonic examination record and image of patient etc.) inspection and image dimension: are shown；
E) doctor's advice dimension: all kinds of doctor's advices of the record doctor to patient；
F) medical history dimension: the electronic health record record of patient；
G) pathology dimension: the pathology of patient is recorded；
H it) nurses dimension: showing the nursing record of patient, such as pulse, body temperature, blood pressure, breathing in graphical form；
I) physical examination dimension: display patient's physical examination record；
J) expense dimension: display statistics all kinds of expense details of patient.
Clinical data mining analysis and aid decision-making method based on internet integration medical platform, including data mining
Analysis and aid decision, data mining analysis include multidimensional analysis algoritic module, data mining algorithm module, deep learning algorithm
Module.Multidimensional analysis algoritic module is firstly the need of cube is established, due to its characteristic with many dimensions, multidimensional data
Collection is usually visually known as data cube (Cube), and cube is a data acquisition system, usually from data warehouse
It is middle to choose several data subsets, then organize and be aggregated into the multidimensional structure as defined in multiple dimensions and metric；Data mining
Algoritic module provides the uniform registration including machine learning algorithms such as classification, cluster, correlation rule, regression analyses, using with
Management is nullified, for being directed to the mining analysis of specific set of data, realizes clinical depth analysis, early warning and prediction；Deep learning
Prediction algorithm module is integrated with the recurrence mind of convolutional neural networks (CNN), Recognition with Recurrent Neural Network (RNN) and shot and long term memory unit
Through network model (LSTM) scheduling algorithm.Aided remote decision by based on index parameter prediction module, be based on audit report text
Four parts such as prediction module, model training module and structurized module composition.
One, multidimensional analysis algoritic module
Multidimensional data analysis is firstly the need of cube is established, due to its characteristic with many dimensions, multidimensional data
Collection is usually visually known as data cube (Cube).Cube is a data acquisition system, usually from data warehouse
It is middle to choose several data subsets, then organize and be aggregated into the multidimensional structure as defined in multiple dimensions and metric.
(1) dimension: refer to that user observes the angle of data, such as hospital is usually concerned about the medical change of number data at any time
Change situation, this is the variation of medical number from coming from the angle of time, and at this moment the time is exactly a dimension.Include dimensional information
Be dimension table, detailed value or factual data comprising saving metric are true tables.Dimension table includes description factual data
The characteristic of true record in table.One dimension is usually with the rank of multiple and different granularities, the i.e. fine degree of viewing angle, example
Such as time dimension can have year, season, the moon different granularity level.
(2) metric: being to have the numerical value of practical significance, such as medical number, drug usage amount etc..Metric is stored in
In true table, true table is the core of analyzed cube, required when being end user's browsing cube
The data checked
Multidimensional data analysis can be carried out upper volume to the data cube organized with multi-dimensional form, lower bore, is sliced, cutting
The a variety of analyses operation such as block, rotation enables analyst, policymaker from multiple angles, multiple side observed numbers so as to profile data
According to the data in library, so that understanding in depth includes information and intension in data.
Two, data mining algorithm module
Data mining algorithm module is provided including machine learning algorithms such as classification, cluster, correlation rule, regression analyses
Uniform registration is managed using with cancellation, for being directed to the mining analysis of specific set of data, realizes clinical depth analysis, early warning
With prediction.
Using data mining algorithm carry out model training, be broadly divided into case study, data prediction, data modeling,
Outcome evaluation, knowledge apply this five steps:
1, case study.By the productive discussions with doctor, expert, the target of explicit data analysis defines data analysis
Demand, setting data analysis expected results.
2, data prediction.Determine data analysis data source, construct reasonable data warehouse for data inquiry with
Analysis.Utilize the databases such as sqlserver, mysql, the high-level languages such as java, python, the statistical means logarithm such as R, spss
It mainly include data integration, data cleansing and data transformation according to necessary pretreatment work is carried out.
3, data modeling.Using processed data as input feature vector, suitable machine learning model is selected, or
Have and is suitably modified on the basis of machine learning model, it is anti-during model training using objective attribute target attribute as output feature
Parameter adjustment and model optimization are carried out again, obtain the optimal training pattern of effect.
4, outcome evaluation.After the completion of model construction, it would be desirable to carry out the assessment of science, to the reliability of model to ensure
The convincingness of experimental result.Specifically, can by the methods of ROC curve and confusion matrix, using accuracy rate, recall rate,
The indexs such as F1 value measure the quality of model.
5, knowledge application.When a model is demonstrated as a reliable, efficient, practical model, a last step
It is that the application of model.All steps that all fronts are done are provided to using preparing, and the application of knowledge is only engineering
Where the core value of habit.Go to predict unknown things using molding model, or seek unknown things and known things it
Between connection, so that people be helped preferably to recognize unknown things.
The classification such as random forest, support vector machines, neural network, decision tree calculation is integrated in data mining algorithm module
Method, K-means clustering algorithm, logistic regression, linear regression and association analysis scheduling algorithm.
(1) the basic classification thought of random forest is by the decision tree obtained after multiple technique drills by bagging
Prediction result is provided by each base classifier, finally takes ballot when inputting a unknown sample to be measured for base classifier
Mode obtain the prediction result of random forest.It is described in detail below: being put firstly, being concentrated with using bagging technology from sample
The subset of the extraction same size returned repeats K times, that is, generates K sample set, then utilizes each self-generating of these sample sets
One corresponding decision tree.Optimal decision tree is generated to each sample set, and knot finally is predicted into the classification of these decision trees
Fruit is voted, classification prediction result of the several most results of getting tickets as Random Forest model.
(2) basic thought of support vector machines is by finding an optimal segmentation hyperplane, so that the plane is being protected
Under the premise of card maximizes classification accuracy, the distance of both sides sample to hyperplane is maximum.Realizing Structural risk minization
On the basis of, generalization ability is promoted, while making every effort to the minimum of empiric risk and confidence interval, accordingly even when less in sample size
In the case where, it can also obtain good classifying quality.
(3) base unit of neural network is known as neuron, it simulates the nerve cell in human brain structure and carries out knowledge
Study, by the way that neuron to be carried out to certain topological sorting, the nerve cell simulated in human brain interconnects and transmits information
Mechanism achievees the purpose that autonomous learning.Neural network can be divided into several component parts in structure, be input layer respectively, hide
Layer and output layer.Wherein input layer receives the initial data of extraneous input, and passes it to hidden layer；Hidden layer may include
One layer or multilayer are responsible for internal information processing and conversion, and the information after conversion are passed to output layer；Information is exporting
Layer is by outwardly exporting final result after last processing and conversion.But under actual conditions, unidirectional TRANSFER MODEL is simultaneously
It cannot be guaranteed that the accuracy of result, needs to adjust the parameter and weight between each layer repeatedly by a large amount of duplicate experiments, with
Seek obtaining preferable training pattern.
(4) regression analysis is a kind of statistical analysis of complementary quantitative relationship between two or more determining variable
Method.According to the quantity for analyzing related variable, it is divided into simple regression analysis and multiple regression analysis；According to independent variable
Quantity, linear regression analysis can be divided into simple regression analysis and multiple regression analysis；According to independent variable and because becoming
Relationship type between amount can be divided into linear regression analysis and nonlinear regression analysis.In regression analysis, if only including one
A independent variable and a dependent variable, and the relationship of the two can approximatively be indicated with straight line, this regression analysis just by
Referred to as simple linear regression analysis.If in regression analysis include two or more independents variable, and dependent variable and independent variable it
Between relationship be it is linear, then it is this analysis be referred to as multiple linear regression analysis.
(5) it whether there is certain dependence between the main data object of correlation analysis, and have dependence for specific
The data object of relationship inquires into its related direction and degree of correlation.Correlativity is a kind of relationship of uncertainty, for example, with
X and Y remembers the situation of change of two kinds of indexs (such as T3 and T4) of a hyperthyroid patient respectively, or is denoted as the height and weight of people,
Then X and Y obviously have relationship, and not definitely to can go accurately to determine another degree by one of those, here it is
Correlativity.
(6) basic thought of Apriori association analysis algorithm: the first step is retrieved in transaction database by iteration
All frequent item sets, i.e. support are more than or equal to the item collection of minimum support set by user；Second step utilizes frequent item set structure
Produce the rule for meeting user's min confidence.Specific practice is exactly: finding out frequent 1- item collection first, is denoted as L1；Then it utilizes
L1 generates candidate C2, carries out decision analysis to the item in C2, excavates L2, i.e., frequent 2- item collection；Constantly so iteration is followed
Ring goes down until can not find more frequently k- item collections.One layer of Lk of every excavation just needs to rescan an entire data
Library.
(7) for K-means algorithm using manhatton distance or Euclidean distance as similarity measure, it is to ask corresponding a certain first
Beginning cluster centre vector v most has classification, so that evaluation index J is minimum.The setting of the K value of K-means clustering algorithm is predefined
, it can be customized by business experience or be obtained by algorithm checks, which represents the number of initial classes cluster centre point,
There is large effect to final cluster result.The algorithm concentrates remaining each object to data in each iteration, according to
Its class that the nearest class heart is re-assigned at a distance from each class center, after having investigated all data objects, once
Interative computation is completed.
(8) temporal sequence association rule dig realize multiple time granularities time it is gauged, such as year, month, day multiple time granularity dimension
The method indicated is spent, is indicated using Linear Segmentation and vector form cluster realizes that the Image Segmentation Methods Based on Features of time series and symbolism turn
The thought changed.In addition when timing is excavated, often some subsequence of time series is excavated, in time series phase
Sliding window is proposed in the application of Time Series Similarity dimensionality reduction technology like the determination of sliding window in property problem.
Three, deep learning prediction algorithm module
Deep learning prediction algorithm module is integrated with convolutional neural networks (CNN), Recognition with Recurrent Neural Network (RNN) and shot and long term
Recurrent neural networks model (LSTM) scheduling algorithm of memory unit.
(1) convolutional neural networks (CNN)
In convolutional neural networks, (it is at C1 layers with the trainable filter fx image inputted that deconvolutes
Input picture, convolutional layer input later are then the convolution characteristic patterns of preceding layer), it (is generally used by an activation primitive
It is Sigmoid function), then plus a biasing bx, convolutional layer Cx is obtained.Concrete operation such as following formula, Mj is input feature vector in formula
The value of figure:
(2) Recognition with Recurrent Neural Network (RNN)
Why RNNs is known as circulation neural network, i.e. the output of a sequence current output and front is also related.Tool
The form of expression of body is that network can remember the information of front and be applied in the calculating currently exported, i.e., between hidden layer
Node it is no longer connectionless but have connection, and not only the output including input layer further includes upper a period of time for the input of hidden layer
Carve the output of hidden layer.Theoretically, RNNs can be handled the sequence data of any length.
(3) recurrent neural networks model (LSTM) of shot and long term memory unit
LSTM is the extension on basic RNN.LSTM is different from the place of RNN, is mainly that it is added in the algorithm
The structure of one " processor " judged whether information is useful, the effect of this processor is referred to as cell.In one cell
Three fan doors have been placed, has been called input gate respectively, forgets door and out gate.One information enters in the network of LSTM, can be with
It is according to rule to determine whether useful.The information for only meeting algorithm certification can just leave, and the information not being inconsistent then passes through forgetting door
It passes into silence.Say be exactly nothing but one-in-and-two-out working principle, can but be solved under operation repeatedly in neural network long-term
Existing big problem.
Aided remote decision: thyroid gland, coronary heart disease and the tumour of selection endocrine subject, cardiovascular subject and tumour subject
Research object etc. several diseases as data collection and analysis relies on unified platform acquisition to integrate clinical medical data, real
The medical diagnosis aid decision-making system towards thyroid nodule, coronary heart disease and tumor of breast etc. is showed, has been examined for clinician's clinic
Disconnected and scientific research personnel's disease research provides system and supports.Mainly by based on index parameter prediction module, based on audit report text
Four parts such as this prediction module, model training module and structurized module composition.
A) based on the prediction module of index parameter: according to the information such as patient's outpatient service serial number or medical insurance card number, Ke Yicha
Ask the test rating and audit report text of the related disease of the patient.Structuring achievement data can be directly inputted；It is right
In non-structured audit report, structuring is carried out using structuring submodule and obtains the data format that model can identify.
B the prediction module) based on audit report text:, can according to the information such as patient's outpatient service serial number or medical insurance card number
To inquire the audit report text of the patient.For non-structured audit report text, deep learning algorithm is directly utilized
It is predicted.
C) model training module: belonging to the basic module of system, invisible to user.By the thyroid gland knot of multiple databases
The data such as the relevant clinical audit reports of the medical information systems such as section, coronary heart disease, mammary gland, test rating merge processing, collect
At into unified tables of data, model training is carried out.
D) structurized module: realize that the structuring to ultrasonic report text data is handled, the ultrasound for extracting various samples is special
Sign includes the index value of Tumor size, boundary, echo distribution, echo intensity etc. and each index, and forms retouching by each sample
State template.Based on the template, the processing of the structuring to ultrasonic content of text is realized.
The present invention selects endocrine subject, cardiovascular subject and the hyperthyroidism of tumour subject, diabetes, thyroid nodule, cream
Several diseases of adenoncus tumor as research object, realize towards hyperthyroidism, diabetes, thyroid nodule, tumor of breast disease disease
Sick clinical data analysis excavates.
Embodiment 1: hyperthyroidism clinical data analysis excavates
Hyperthyroidism clinical data analysis excavate realize to the clinical data of hyperthyroidism disease carry out multidimensional analysis, association analysis and
Clustering etc..True clinical medical data of the hyperthyroidism clinical data from multicenter clinic big data platform.Hyperthyroidism clinic number
According to the general frame of analysis mining as shown in Fig. 2, be largely divided into four levels, be respectively data active layer, data Layer, analysis layer and
Application layer.The data put in order are loaded among HANA, later from raw data set by preprocessing process such as ETL by data
Relevant analysis purpose is realized in conjunction with correlation analysis algorithm using analysis tool.
Data source: this level corresponds to initial data, is the major database of structured data.It contains basic
The information such as information table, diagnosis records table, survey report table, test rating table, diagnostics table, prescription detail list.This layer is thereon
The operation basis of face all levels, it is necessary to assure work normally.
Data preparation layer: this level is responsible for integrate from the data cleansing of data source, constructs and meets multidimensional point
The data set of analysis model will extract the data acquisition system excavated simultaneously for association rule mining later.Below
Whether data analysis and the result excavated are accurate, whether thorough, the number of building of the data cleansing for this level that places one's entire reliance upon
Whether meet the requirements according to set.
Data Layer: this level is the realization of Data Analysis Model and mining model.I.e. by ready data, there are this
Among layer, pending datas analysis and tap layer is waited to operate on it.
Analysis layer: analysis layer is the core level of whole system, including multidimensional analysis and the big module of association rule mining two.
Multidimensional analysis module designs suitable Star Model, to complete the analysis work of specific subject according to analysis theme；Association
Rule digging module is then in alignment with the associated data set got ready and does specific excavation, if data set owner according to the theme of excavation and
It is fixed.
Application layer: the processing of layer by analysis can generate experimental result, how clearly show result, with regard to this
The work of this level.Result and digging are analyzed using the WebI component of SAP and the visualization analysis tools of HANA to show respectively
Dig result.
The level of system is divided into three big modules, first be initial data preprocessing module, which passes through to clinic
Tables of data carries out cleaning, obtains the clinical medical data for meeting specification, completes data cube and associated data again later
The building task of collection；Second module is data analysis module, which will complete base according to the multidimensional analysis models defined
Three eigen analysis, analysis of drug use and index analysis subtasks；Third is data-mining module, this module will mainly close
Connection rule mining algorithms are applied in associated data set, obtain the result set of correlation rule.It is last then using WebI and HANA
Visualization technique shows result.
Embodiment 2: diabetes clinical data analysis and excavation
Diabetes clinical data analysis, which excavates to realize, carries out multidimensional analysis, correlation analysis to the clinical data of diabetes
With diagnostic event timing excavate etc..True clinical medical data of the diabetes clinical data from integral system platform.
According to the function of diabetes clinical data analysis application system, Fig. 3 is the hierarchical chart of system, mainly there is three
A functional module.First module is data preprocessing module, because may require that different structure for different analysis demands
Data carry out data needed for processing obtains multidimensional analysis and Time-Series analysis to clinical diabetes database；Second module be
Diabetes multidimensional analysis module, multidimensional analysis are one of the nucleus module for analyzing process layer, the multidimensional analysis to diabetes data
The feature (such as: age, gender, area) of diabetic patient population can be observed according to diagnosis, index and medicining condition；Third
A module is that timing excavates module, and timing excavation, which is analysis another nucleus module of process layer, to be often accompanied by for diabetes
The reality of more complication excavates a timing of the complication occurrence regularity of clinical diabetes.
Diabetes clinical data analysis application system hierarchical chart is followed successively by data active layer, ETL layers, analysis from top to bottom
Next process layer and application layer briefly introduce basic function at all levels.
1. data active layer: being the major database for storing clinical structural data.Contain Basic Information Table, visiting hospital register
Information table, clinical diagnosis table, prescription detail list, clinical examination index table etc..This layer is the basis that system operates normally, and
Post analysis work initial data.
2. data preparation layer: in data analysis, the data structure that different analysis methods generally requires is also different, this
Process layer is analyzed in text, and mainly there are two analysis directions: the Time-Series analysis of multidimensional analysis and diagnostic event.So in data preparation
Layer is divided into the processing method of three subdivisions again, is data cleansing, building multi-dimension data cube, building diagnostic event sequence respectively.
3. analyzing process layer: analysis process layer is the core level of diabetes clinical data analysis application system, layer master
Will be there are two functional module, the Time-Series analysis module of multidimensional analysis module and diagnostic event respectively corresponds on-line analytical processing sum number
According to excacation.Multidimensional analysis module designs Star Model according to analysis theme, Data Mart is constructed, thus analysis of diabetes
The essential characteristic of PATIENT POPULATION；Time-Series analysis module examines patient in the primary medical diagnosis of hospital as one of the patient
Disconnected event just has time upper successive concept between such diagnostic event, and each patient's body is exactly a sequence, finally to true
A sequence sets carry out the excavation of frequent mode, obtain the frequent mode for meeting minimum support.
4. applying presentation layer: by analyzing process layer, mainly carrying out probe into application in terms of two herein, be multidimensional respectively
Analysis and Time-Series analysis.Multidimensional analysis module is mainly corresponding with analysis of drug use, index analysis and diabetes and its complication
Diagnostic analysis；Time-Series analysis module mainly to analysis of experimental results, is tied experiment from the setting of minimum support and confidence level
Fruit analysis.The result of analysis can be presented in system in a manner of chart, report etc..
It can be three main module compositions according to the hierarchical chart of system.First be data preparation layer data
Preprocessing module, the module mainly have diagnostic event sequence three data cleansing, building multi-dimension data cube, building operations.Second
A module is to analyze the multidimensional analysis module of process layer, according to analysis demand tissue multidimensional data, mainly there is analysis of drug use, index
Analysis and three parts of clinical analysis of diagnosis.Third module is to analyze the timing excavation module of process layer, proposes NFPS frequency
Numerous mode discovery algorithm.Finally, use intuitiveization of the result of multidimensional analysis is (such as: chart, datagram to analysis result visualization
Table etc.) form presentation come out.
Main function of system is divided into two lines, is multidimensional analysis and time series analysis respectively, finally carrying out to result can
Depending on changing.Firstly, system first pre-processes the data of acquisition, pretreatment does different processing according to the difference of analysis demand,
Then multidimensional analysis and the Time-Series analysis of diagnostic event are respectively enterd, finally by the result pictorialization of multidimensional analysis, and to timing
The algorithm of excavation compares experiment and carries out correlation analysis to the reliability of result set.
During analyzing clinical diabetes diagnosis, discovery diabetes are frequently accompanied by many complication, Er Qie
There are different complication in the different state of an illness stages, whether there is certain association between diabetic complication to explore.By
There is no the concept of time series in traditional Multi-relational frequent pattern discovery, in order to find successive association between event, is faced according to diabetes
The characteristics of bed diagnostic data set, Multi-relational frequent pattern discovery Time-Series analysis is carried out, NFPS algorithm is proposed.The step of algorithm, describes such as
Under:
It inputs diagnostic event sequence sets D and time window constrains G；
1) Effect-Sequence algorithm is carried out, the sequence sets D ' for meeting time window constraint is obtained
2) ergodic data collection calculates a frequent item collection
3) it is connected by K item collection and obtains K+1 candidate
4) if carrying out cut operator // K item collection according to minimum support is frequently, then all items of K-1 item collection also must
It must be frequent
5) Recursive Implementation step 3) and the item collection 4) until not expiring minimum support
6) the fuzzy frequent itemsets S ' of diagnostic event sequence is obtained
From above-mentioned steps as can be seen that the algorithm mainly consists of three parts, the validation of a sequence sets is carried out first,
What this step executed is actually Effect-Sequence algorithm, obtains candidate frequent episode from connection followed by item collection；Finally
It is beta pruning part, can all has cut operator in recursive each step, if it is frequent that the foundation of cut operator, which is K item collection, then
The K-1 item collection for forming K item collection also must all be frequent.
Embodiment 3: thyroid nodule clinical data analysis excavates
Thyroid nodule clinical data analysis, which excavates to realize, carries out multidimensional analysis, disease to the clinical data of thyroid nodule
Disease classification and good pernicious differentiation etc..True clinical medical data of the thyroid nodule clinical data from clinical data platform.
Fig. 4 is that thyroid nodule clinical data analysis excavates integrated stand composition, can be seen that thyroid nodule is clinical in figure
Data analysis mining is mainly made of input module, training module and prediction and display module three parts, wherein data prediction
The basic module of system is partly belonged to, it is invisible to user.User provides initial data, is obtained using microstructured tool or technology
Then the data format that model can identify carries out model construction using machine learning algorithm, best model is selected to carry out unknown sample
This diagnostic classification prediction, finally shows user to consult result.
The data of multiple files or multiple databases are mainly merged processing by data integration, relate generally to data
Selection, data conflict and data it is inconsistent the problems such as processing problem.During data integration, need to consider field
Definition, the selection of data type etc..The data of thyroid disease are more dispersed, are distributed in different database tables, number
Effectively these data can be integrated according to integrating process.For example, extracting the basic of patient from patient basis's table
Information such as gender, age etc.；Index name, index value etc. are extracted from Index for examination table；Diagnosis name etc. is extracted from diagnostics table,
By completing to be integrated into unified tables of data after extracting to aforesaid operations.
It finds in practice, extracting related data by multiple database tables and integrating to a database table, treated counts
It is many according to missing values, it needs to be further processed.By serial number of going to a doctor, the inspection record extraction of patient is handled, and is integrated into
One record.There are missing values for T3 the and T4 index of most of patient, it is necessary to be handled in order to avoid influencing subsequent analysis result.This
It is literary mainly to take the method deleted missing column and replace missing values using regression analysis.For the category containing most missing values
Property, it is necessary to it deletes.It, can be with completion missing values for containing the sample of a small amount of missing values.
The quantity that medical system generates data is big, process is complicated, repetition, missing even mistake of data etc. inevitably occurs,
In order to reduce the interference of these noises, accuracy when improving followed by classification can be by there is the data cleansing of supervision calculation
Method obtains effective data.
Data cleansing mainly includes the noise data and extraneous data removed in original data set, and processing missing values and cleaning are dirty
Data etc., and complete some data type conversion work.The data cleansing process of this thyroid gland medical data is medical special
Under family's guidance, the data after integrated are analyzed and processed, noise data and Data duplication record is removed, fills up missing data.
Generally for the processing method of missing values, following method can be taken according to different situations:
(1) missing values are deleted: whether including key message according to each record to determine reservation or deletes the note
Record.If the attribute value of a record missing is too many, or not includes key message, such as the card number or medical stream of patient
When water number lacks, the record is deleted herein.
(2) constant value method of substitution: the attribute value of all missings is filled with the same constant such as NULL, and this method is very simple
It is single, it is mainly used for the processing of the fields such as the marital status in patient's Basic Information Table.
(3) when field is numeric type data, all missings mean value method of substitution: can be filled with the average value of the attribute
Value.Such as in patient's Basic Information Table, the age of a small number of patients without the date of birth takes the average value of all patient ages.
(4) estimated value method of substitution: the predicted value of the attribute missing values is obtained with the methods of regression analysis, is filled with it scarce
Mistake value.Such as the missing values of T3 and T4 are calculated by regression analysis in test rating table.It is used during executing data cleansing
Which type of processing method will be determined according to the concrete condition of data set and related request.Regression analysis pair is utilized herein
Missing values are handled.
Data conversion is mainly to reduce the number of useful variable or find the invariants of data, including normalization, conclude, cut
The operation such as change, rotate and project.Computational efficiency can be greatly improved by data transformation, and the starting point of Knowledge Discovery can be improved.
For example for Gender, male's attribute value can be set as to 0, corresponding women attribute value is set as 1.Diagnostic result point in diagnostics table
Abnormal higher to be normal, corresponding attribute value can be used 1,2 and 3 replacement by abnormal relatively low three kinds of situations respectively.
The disaggregated model of thyroid disease
For the clinical data of thyroid disease, a kind of classification side of thyroid disease type based on random forest is proposed
Method, this method use Principal Component Analysis to carry out feature selecting to data set first, reduce data dimension, then using random gloomy
Woods algorithm realizes classification.The model mainly analyzes many index data set in the serum of patient using data mining, from
And realize the classification of thyroid disease, it is mainly made of three phases, method flow is shown in Fig. 5.
First stage: data prediction.
Firstly, it is necessary to establish disaggregated model by training set.In actual inspection, due to the 5 of thyroid function inspection
Index is not essential items for inspection, and there are a large amount of missing values in the medical data base of hospital, so needing to carry out data prediction.I
The object that extracts be the patient that thyroid disease is diagnosed as in diagnostics table, by patient's Basic Information Table, inspection result index table
It is merged with the data of multiple tables such as diagnostics table by the way that multilist is operation associated, by selection, transformation, a series of ETL such as integrates
Operation obtains related data, establishes the thyroid gland integrated data set comprising patient's essential information and test rating.According to medical stream
Water number is by the data preparation of each patient at including gender, age, the total triiodo thryonine of serum (T3), thyroxine
(T4), free serum triiodothyronine (FT3), free thyroxine (FT4), the attributes such as thyrotropic hormone (TSH)
A plurality of record.
Second stage: Feature Dimension Reduction.
Feature selecting (Feature Selection) is also referred to as Attributions selection (Attribute Selection), refers to root
According to certain criterion from known feature set, it is advantageously selected for distinguishing the character subset of different classes of data.Feature selecting can pick
Except some incoherent features, Characteristic Number is reduced, raising model is accurate, reduces runing time.
In order to retain the main information of initial data, dimension-reduction treatment is carried out to data using Principal Component Analysis herein.It will
Thyroid gland data characteristics collection F random division is at k subset Fij, FijIt indicates for training classifier DiJ-th of character subset.It takes out
The sample of each subset 75% is taken to establish new subset, it is therefore an objective to improve the otherness of base classifier.To new thyroid gland data
Character subset carries out feature selecting, reduces attribute dimensions.
Principal Component Analysis is a kind of statistical method of dimensionality reduction, its effect can be reduced comprising a large amount of association attributes numbers
According to the dimension of collection.In order to extract main information from multidimensional data, the primitive attribute space of thyroid gland data set is transformed into category
The incoherent new space of property.The principal component that original variable is identified by linear combination, defines original thyroid gland data set
Maximum middle variance is first principal component, is Second principal component, in remaining data set.Because variance accumulative perception has letter
Unisexuality and qualified performance are so for determining the number of principal component.Accumulative variance percentage is usually between 70% and 90%
Range carries out the selection of principal component when higher than defined threshold value.
Enabling each attribute of thyroid gland data is xi, then
Ci=ai1x1+ai2x2+······+aipxp, i=1,2 ...,
Wherein, x1,x2,...,xpFor stochastic variable, aijReferred to as principal component coefficient, if Var (Ci) is maximum, Ci is referred to as
First principal component.Similarly, can have second, third, the 4th ... principal component, at most have p.
M is enabled to represent all principal component numbers, p indicates most important principal component quantity in principal component, and p is that m ties up thyroid gland
The quantity of principal component in data with highest variance yields, by analyzing visible p≤m, it is achieved that original thyroid gland data
The dimensionality reduction of collection.
Phase III: classification prediction.
The selection of base classifier can have large effect to the precision of Ensemble classifier algorithm, in order to improve thyroid gland disease
The precision of disease classification, compared common base classifier algorithm such as Naive Bayes, SMO, C4.5 on thyroid gland data set
Performance.
Since there are large amount of text information for Thyroid ultrasound index, examined using a kind of pathology based on interdependent syntactic analysis
Report structure method is looked into, detailed process is as follows: firstly, for a variety of description feelings of same index frequently occurred in pathological replacement
Condition is pre-processed, and finds out term vector using neural network model, is calculated cosine similarity on this basis and is found out synonym, advises
The text expression of model pathologic finding report, while cutting short sentence and introducing word information labeling method and simplify a sentence structure, it reduces
The height of dependency tree improves the accuracy of structured result so that grammatical relation be made to be more clear；Followed by interdependent sentence
Method is analyzed to obtain the dependency tree of each short sentence, extracts index and corresponding index using gained semantic feature and part of speech feature
Non-structured text, can be converted to the structured stencil of key-value form by value；Finally markup information is restored, simultaneously
Correct noise data.According to the difference for realizing function, total process can be divided into 3 modules: preprocessing module, knot
Structure module, post-processing module.
Embodiment 4: tumor of breast clinical data analysis excavates
The excavation of tumor of breast clinical data analysis, which is realized, to be associated analysis to the clinical data of tumor of breast disease, gathers
Alanysis and classification analysis etc..True clinical medical data of the tumor of breast clinical data from clinical big data platform.
Tumor of breast Data Mart mainly includes following five base tables: mammary gland patient checks summary table, breast X-ray report
Table, breast ultrasound account, breast puncture result table and mammary gland pathological account.
Breast ultrasound diagnostic result is divided into different brackets, and the final pathological diagnosis result of patient is divided into benign with pernicious two
Kind.
It is proposed a kind of new knowledge mapping inference method based on TransR-DNN, the mammary gland for constructing high anticipation accuracy rate is swollen
The good pernicious discrimination model of tumor.According to the tumor of breast knowledge mapping that completion is perfect, a kind of new knowledge based map is proposed
Inference Forecast algorithm, carry out Breast Tumors prediction differentiate.Tumor of breast clinical fact knowledge graph is analyzed first
Modal data amount is big, and comprising abundant semantic space, and there are the relationships of a large amount of multi-to-multis, herein based on translation conversion TransR
On model, new model TransR-DNN learning algorithm is proposed, by predicting link and the entity of triple, to obtain accurately
The higher Breast Tumors of rate predict discrimination model.New model finally is assessed from accuracy, recall rate and F1 score, and
And comparative experiments is carried out on time complexity and time loss, it is more excellent to demonstrate new model performance.
In conclusion using technical solution of the present invention, the clinical data based on internet integration medical platform is dug
Pick analysis and aid decision-making method, unstructured clinical document are input to clinical document structuring processing engine, are cured by clinic
The means processing such as corpus, rule, full-text search and machine learning is learned, obtains structural data and be output to distributed storage drawing
It holds up, is handled by intelligent algorithm, for Platform Analysis, shown；The present invention is by text non-structured in clinical data
Notebook data carries out structuring processing, stores into distributed Hadoop cluster, realizes Distributed Storage mode and distribution
Calculation processing, and the programming in software application is realized and is transformed and is adapted to for distributed nature.
Claims (6)
1. clinical data mining analysis and aid decision-making method based on internet integration medical platform, it is characterised in that: packet
Data mining analysis and aid decision are included, data mining analysis includes multidimensional analysis algoritic module, data mining algorithm module, depth
Spend learning algorithm module；Multidimensional analysis algoritic module is firstly the need of cube is established, and cube is from data warehouse
Several data subsets are chosen, then organize and be aggregated into the multidimensional structure as defined in multiple dimensions and metric；Data mining is calculated
Method module provides the uniform registration including machine learning algorithms such as classification, cluster, correlation rule, regression analyses, using with note
Pin pipe reason realizes clinical depth analysis, early warning and prediction for being directed to the mining analysis of specific set of data；Deep learning is pre-
Method of determining and calculating module is integrated with the recurrent neural of convolutional neural networks (CNN), Recognition with Recurrent Neural Network (RNN) and shot and long term memory unit
Network model (LSTM) scheduling algorithm；Aided remote decision by based on index parameter prediction module, based on audit report text
Prediction module, model training module and structurized module composition.
2. clinical data mining analysis and aid decision side as described in claim 1 based on internet integration medical platform
Method, which is characterized in that the multidimensional analysis algoritic module to the data cube organized with multi-dimensional form carry out upper volume, under
Brill, slice, stripping and slicing, a variety of analyses of rotation operate, and so as to profile data, enable analyst, policymaker from multiple angles, multiple sides
The data in database are observed in face, so that understanding in depth includes information and intension in data.
3. clinical data mining analysis and aid decision side as claimed in claim 2 based on internet integration medical platform
Method, which is characterized in that the multidimensional analysis algoritic module includes dimension and metric, and the dimension is the angle that user observes data
Degree, what it is comprising dimensional information is dimension table, and detailed value or factual data comprising preservation metric are true table, dimension table packet
The characteristic of true record in the fact table containing description；The metric is the numerical value with practical significance, metric storage
In true table.
4. clinical data mining analysis and aid decision side as claimed in claim 3 based on internet integration medical platform
Method, which is characterized in that it is described using data mining algorithm carry out model training, be broadly divided into case study, data prediction,
Data modeling, outcome evaluation, knowledge apply five steps:
(1) case study: by the productive discussions with doctor, expert, the target of explicit data analysis defines the need of data analysis
It asks, the expected results of setting data analysis；
(2) data prediction: determining the data source of data analysis, constructs reasonable data warehouse for the inquiry of data and divides
Analysis, using sqlserver, mysql database, java, python high-level language, R, spss statistical means must to data progress
The pretreatment work wanted；
(3) data modeling: using processed data as input feature vector, suitable machine learning model is selected, or existing
It is suitably modified on the basis of machine learning model, using objective attribute target attribute as output feature, during model training repeatedly
Parameter adjustment and model optimization are carried out, the optimal training pattern of effect is obtained；
(4) outcome evaluation: after the completion of model construction, it would be desirable to carry out the assessment of science, to the reliability of model to ensure reality
The convincingness of result is tested, specifically, accuracy rate, recall rate, F1 can be utilized by the methods of ROC curve and confusion matrix
The indexs such as value measure the quality of model；
(5) it knowledge application: goes to predict unknown things using molding model, or seeks between unknown things and known things
Connection, so that people be helped preferably to recognize unknown things.
5. clinical data mining analysis and aid decision side as claimed in claim 4 based on internet integration medical platform
Method, which is characterized in that be integrated with random forest in the data mining algorithm module, support vector machines, neural network, return and divide
Analysis, correlation analysis, Apriori association analysis, K-means clustering algorithm；
(1) random forests algorithm is described in detail below: firstly, being concentrated with the extraction phase put back to from sample using bagging technology
With the subset of size, repeats K times, that is, generate K sample set, it is then corresponding using each self-generating of these sample sets one
Decision tree generates optimal decision tree to each sample set, and finally the classification prediction result of these decision trees is voted,
Classification prediction result of the several most results of getting tickets as Random Forest model；
(2) algorithm of support vector machine finds an optimal segmentation hyperplane, so that the plane is guaranteeing to maximize classification accurately
Under the premise of rate, the distance of both sides sample to hyperplane is maximum；
(3) neural network algorithm can be divided into several component parts in structure, be input layer, hidden layer and output layer respectively,
Middle input layer receives the initial data of extraneous input, and passes it to hidden layer；Hidden layer may include one layer or multilayer,
It is responsible for internal information processing and conversion, and the information after conversion is passed into output layer；Information is in output layer by last
Processing outwardly exports final result with after conversion；
(4) algorithm with regress analysis method is divided into simple regression analysis and multiple regression according to the quantity for analyzing related variable
Analysis；According to the quantity of independent variable, linear regression analysis can be divided into simple regression analysis and multiple regression analysis；It presses
According to the relationship type between independent variable and dependent variable, linear regression analysis and nonlinear regression analysis can be divided into；
(5) it whether there is certain dependence between correlation analysis algorithm data object, and have dependence for specific
Data object inquire into its related direction and degree of correlation；
(6) Apriori association analysis algorithm, the first step retrieve all frequent item sets in transaction database by iteration；The
Two steps construct the rule for meeting user's min confidence using frequent item set；
(7) for K-means algorithm using manhatton distance or Euclidean distance as similarity measure, it is to ask corresponding a certain initial poly-
Class center vector v most has classification, so that evaluation index J is minimum.
6. clinical data mining analysis and aid decision side as described in claim 1 based on internet integration medical platform
Method, which is characterized in that the prediction module based on index parameter: according to patient's outpatient service serial number or medical insurance card number information,
The test rating and audit report text for inquiring the related disease of the patient, can directly input structuring achievement data；
For non-structured audit report, structuring is carried out using structuring submodule and obtains the data format that model can identify；
The prediction module based on audit report text: according to the information such as patient's outpatient service serial number or medical insurance card number, inquiry
To the audit report text of the patient；For non-structured audit report text, directly carried out using deep learning algorithm pre-
It surveys；
The model training module: by the relevant clinical audit report of multiple database medical information systems, test rating data
Processing is merged, is integrated into unified tables of data, model training is carried out；
The structurized module: realizing that the structuring to ultrasonic report text data is handled, extract the ultrasonic feature of various samples,
And the index value of each index, and form the description template by each sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910101985.XA CN109830303A (en) | 2019-02-01 | 2019-02-01 | Clinical data mining analysis and aid decision-making method based on internet integration medical platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910101985.XA CN109830303A (en) | 2019-02-01 | 2019-02-01 | Clinical data mining analysis and aid decision-making method based on internet integration medical platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109830303A true CN109830303A (en) | 2019-05-31 |
Family
ID=66863183
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910101985.XA Pending CN109830303A (en) | 2019-02-01 | 2019-02-01 | Clinical data mining analysis and aid decision-making method based on internet integration medical platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109830303A (en) |
Cited By (79)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110277172A (en) * | 2019-06-27 | 2019-09-24 | 齐鲁工业大学 | A kind of clinical application behavior analysis system and its working method based on efficient negative sequence mining mode |
| CN110415831A (en) * | 2019-07-18 | 2019-11-05 | 天宜(天津)信息科技有限公司 | A kind of medical treatment big data cloud service analysis platform |
| CN110477954A (en) * | 2019-07-08 | 2019-11-22 | 无锡海斯凯尔医学技术有限公司 | Detection device based on elastogram |
| CN110544528A (en) * | 2019-08-29 | 2019-12-06 | 中南大学 | advanced learning-based upper and lower ophthalmic remote diagnosis platform and construction method thereof |
| CN110569372A (en) * | 2019-09-20 | 2019-12-13 | 四川大学 | construction method of heart disease big data knowledge graph system |
| CN110570921A (en) * | 2019-08-20 | 2019-12-13 | 广东省第二中医院（广东省中医药工程技术研究院） | Clinical information processing system based on single disease category |
| CN110675952A (en) * | 2019-09-19 | 2020-01-10 | 上海腾程医学科技信息有限公司 | Checking decision method and device, terminal equipment and computer readable storage medium |
| CN110737731A (en) * | 2019-10-25 | 2020-01-31 | 徐州工程学院 | accumulation fund user data refinement analysis system and method based on decision tree |
| CN110767317A (en) * | 2019-08-30 | 2020-02-07 | 贵州力创科技发展有限公司 | Cloud computing platform and method based on data mining and big data analysis |
| CN110875095A (en) * | 2019-09-27 | 2020-03-10 | 长沙瀚云信息科技有限公司 | Standardized clinical big data center system |
| CN111081016A (en) * | 2019-12-18 | 2020-04-28 | 北京航空航天大学 | Urban traffic abnormity identification method based on complex network theory |
| CN111341456A (en) * | 2020-02-21 | 2020-06-26 | 中南大学湘雅医院 | Method and device for generating diabetic foot knowledge map and readable storage medium |
| CN111382151A (en) * | 2020-03-23 | 2020-07-07 | 新疆大学 | CT medical image cleaning method based on data desensitization |
| CN111428930A (en) * | 2020-03-24 | 2020-07-17 | 中电药明数据科技(成都)有限公司 | GBDT-based medicine patient using number prediction method and system |
| CN111696666A (en) * | 2020-06-10 | 2020-09-22 | 杭州联众医疗科技股份有限公司 | Intelligent chronic disease management system based on time coding |
| CN111696669A (en) * | 2020-06-15 | 2020-09-22 | 山东搜搜中医信息科技有限公司 | Quantitative diagnosis of clinical multidimensional data of traditional Chinese medicine |
| CN111696665A (en) * | 2020-06-10 | 2020-09-22 | 杭州联众医疗科技股份有限公司 | Auxiliary decision making system based on time coding |
| CN111696675A (en) * | 2020-05-22 | 2020-09-22 | 平安国际智慧城市科技股份有限公司 | User data classification method and device based on Internet of things data and computer equipment |
| CN111710427A (en) * | 2020-06-17 | 2020-09-25 | 广州市金域转化医学研究院有限公司 | Cervical precancerous early lesion stage diagnosis model and establishment method |
| CN111724898A (en) * | 2020-06-15 | 2020-09-29 | 中国医学科学院医学信息研究所 | Intelligent skin disease monitoring and early warning system based on big data technology |
| CN111739639A (en) * | 2020-06-19 | 2020-10-02 | 杭州联众医疗科技股份有限公司 | Perioperative complication data prediction system based on time coding |
| CN111768846A (en) * | 2020-05-27 | 2020-10-13 | 医利捷(上海)信息科技有限公司 | Clinical data management method |
| CN111863267A (en) * | 2020-07-08 | 2020-10-30 | 首都医科大学附属北京天坛医院 | Data information acquisition method, data analysis device and storage medium |
| CN111899828A (en) * | 2020-07-31 | 2020-11-06 | 青岛百洋智能科技股份有限公司 | Knowledge graph driven breast cancer diagnosis and treatment scheme recommendation system |
| CN111914026A (en) * | 2020-07-31 | 2020-11-10 | 南京朗赢信息技术有限公司 | General data exchange sharing service platform |
| CN111949801A (en) * | 2020-07-27 | 2020-11-17 | 西北工业大学 | Knowledge graph fusion method of doctor experience knowledge and ultrasonic image information |
| CN111951976A (en) * | 2020-08-21 | 2020-11-17 | 上海交通大学医学院附属第九人民医院 | Value judgment method, system, terminal and medium based on medical data margin |
| CN111984987A (en) * | 2020-09-01 | 2020-11-24 | 上海梅斯医药科技有限公司 | Method, device, system and medium for desensitization and reduction of electronic medical record |
| WO2020233254A1 (en) * | 2019-07-12 | 2020-11-26 | 之江实验室 | Medical data analysis system integrating structured image data |
| CN112037925A (en) * | 2020-07-29 | 2020-12-04 | 郑州大学第一附属医院 | LSTM algorithm-based early warning method for newly-released major infectious diseases |
| CN112100286A (en) * | 2020-08-14 | 2020-12-18 | 华南理工大学 | Computer-aided decision-making method, device and system based on multi-dimensional data and server |
| CN112150209A (en) * | 2020-06-19 | 2020-12-29 | 南京理工大学 | Construction method of CNN-LSTM time sequence prediction model based on clustering center |
| CN112185585A (en) * | 2020-11-03 | 2021-01-05 | 浙江大学滨海产业技术研究院 | Diabetes early warning method based on metabonomics |
| CN112200374A (en) * | 2020-10-15 | 2021-01-08 | 平安国际智慧城市科技股份有限公司 | Medical data processing method, device, electronic equipment and medium |
| CN112286985A (en) * | 2020-10-13 | 2021-01-29 | 江苏云脑数据科技有限公司 | Clinical research statistical analysis system based on cloud computing |
| CN112365976A (en) * | 2020-11-14 | 2021-02-12 | 南昌大学第二附属医院 | Compound disease clinical path construction method and system based on transfer learning |
| CN112380763A (en) * | 2020-11-03 | 2021-02-19 | 浙大城市学院 | System and method for analyzing reliability of in-pile component based on data mining |
| CN112446862A (en) * | 2020-11-25 | 2021-03-05 | 北京医准智能科技有限公司 | Dynamic breast ultrasound video full-focus real-time detection and segmentation device and system based on artificial intelligence and image processing method |
| TWI723868B (en) * | 2019-06-26 | 2021-04-01 | 義守大學 | Method for applying a label made after sampling to neural network training model |
| CN112635074A (en) * | 2020-12-21 | 2021-04-09 | 云南省疾病预防控制中心 | AIDS prevention and treatment decision method based on multi-data analysis model |
| CN112786126A (en) * | 2020-12-31 | 2021-05-11 | 天津开心生活科技有限公司 | Time sequence analysis method and device of clinical test data, electronic equipment and medium |
| CN112927810A (en) * | 2021-03-23 | 2021-06-08 | 崔剑虹 | Smart medical response method based on big data and smart medical cloud computing system |
| CN112988783A (en) * | 2021-03-12 | 2021-06-18 | 李涛 | Public opinion occurrence time sequence analysis method based on multidimensional data model |
| CN113053479A (en) * | 2019-12-27 | 2021-06-29 | 天津幸福生命科技有限公司 | Medical data processing method, device, medium and electronic equipment |
| CN113096817A (en) * | 2021-04-13 | 2021-07-09 | 北京大学 | Method, apparatus, computer device and storage medium for disease prediction |
| CN113160999A (en) * | 2021-04-25 | 2021-07-23 | 厦门拜特信息科技有限公司 | Data structured analysis system and data processing method for medical decision |
| CN113177040A (en) * | 2021-04-29 | 2021-07-27 | 东北大学 | Full-process big data cleaning and analyzing method for aluminum/copper plate strip production |
| CN113314201A (en) * | 2021-06-17 | 2021-08-27 | 南通市第一人民医院 | Neurology clinical nursing potential safety hazard analysis method and system |
| WO2021175038A1 (en) * | 2020-11-13 | 2021-09-10 | 之江实验室 | Patient data visualization method and system for assisting decision-making in chronic disease |
| CN113380360A (en) * | 2021-06-07 | 2021-09-10 | 厦门大学 | Similar medical record retrieval method and system based on multi-mode medical record map |
| CN113436747A (en) * | 2021-07-20 | 2021-09-24 | 四川省医学科学院·四川省人民医院 | Medical data clinical auxiliary system and method based on medical data analysis model |
| CN113436745A (en) * | 2021-06-30 | 2021-09-24 | 四川大学华西医院 | Artificial intelligence auxiliary diagnosis method based on database analysis |
| CN113539471A (en) * | 2021-03-26 | 2021-10-22 | 内蒙古卫数数据科技有限公司 | Auxiliary diagnosis method and system for hyperplasia of mammary glands based on conventional inspection data |
| CN113539414A (en) * | 2021-07-30 | 2021-10-22 | 中电药明数据科技(成都)有限公司 | Method and system for predicting rationality of antibiotic medication |
| CN113611411A (en) * | 2021-10-09 | 2021-11-05 | 浙江大学 | Body examination aid decision-making system based on false negative sample identification |
| CN113609555A (en) * | 2021-07-16 | 2021-11-05 | 黄河勘测规划设计研究院有限公司 | Hydraulic metal structure design method based on big data technology |
| CN113628734A (en) * | 2021-08-02 | 2021-11-09 | 浙江海心智惠科技有限公司 | Design method of oncology electronic medical advice system with clinical decision intelligent recommendation function |
| CN113674867A (en) * | 2021-07-27 | 2021-11-19 | 上海药慧信息技术有限公司 | Clinical data mining method and device, electronic equipment and storage medium |
| CN113688169A (en) * | 2021-08-11 | 2021-11-23 | 北京科技大学 | Mine potential safety hazard identification and early warning system based on big data analysis |
| CN113744873A (en) * | 2021-11-08 | 2021-12-03 | 浙江大学 | Heating to-be-checked auxiliary differential diagnosis system based on task decomposition strategy |
| CN113823421A (en) * | 2021-08-20 | 2021-12-21 | 武汉心络科技有限公司 | Information providing method, device, equipment and storage medium |
| CN113889279A (en) * | 2021-09-28 | 2022-01-04 | 北京华彬立成科技有限公司 | Combination therapy information mining and inquiring method, device and electronic equipment |
| CN114068013A (en) * | 2021-11-16 | 2022-02-18 | 高峰 | Cerebral artery occlusion artificial intelligence assistant decision system |
| CN114121293A (en) * | 2021-11-12 | 2022-03-01 | 北京华彬立成科技有限公司 | Clinical trial information mining and inquiring method and device |
| CN114464314A (en) * | 2022-02-08 | 2022-05-10 | 四川大学华西医院 | Clinical body symptom classification diagnosis system |
| CN114496177A (en) * | 2022-01-24 | 2022-05-13 | 佳木斯大学 | Method and system for detecting clinical infection source of infectious department based on big data |
| CN114564755A (en) * | 2022-03-03 | 2022-05-31 | 曾迎春 | Cancer data management platform based on block chain technology |
| CN114678132A (en) * | 2022-02-22 | 2022-06-28 | 北京颐圣智能科技有限公司 | Self-learning medical wind control system and method based on clinical behavior feedback |
| CN114912804A (en) * | 2022-05-17 | 2022-08-16 | 四川大学华西医院 | Scientific research data related property control method and system |
| CN115083601A (en) * | 2022-07-25 | 2022-09-20 | 四川省医学科学院·四川省人民医院 | Type 2diabetes auxiliary decision making system based on machine learning |
| CN115083616A (en) * | 2022-08-16 | 2022-09-20 | 之江实验室 | Chronic nephropathy subtype mining system based on self-supervision graph clustering |
| CN115145993A (en) * | 2022-07-05 | 2022-10-04 | 西南交通大学 | Railway freight big data visualization display platform based on self-learning rule operation |
| CN115240800A (en) * | 2022-09-26 | 2022-10-25 | 北京泽桥医疗科技股份有限公司 | Medical data intelligent analysis execution method based on big data platform |
| WO2022228473A1 (en) * | 2021-04-27 | 2022-11-03 | 联峰远程健康管理服务有限公司 | Smart health management system for use in telemedicine service and method used in same |
| CN115617840A (en) * | 2022-12-19 | 2023-01-17 | 江西曼荼罗软件有限公司 | Medical data retrieval platform construction method, system, computer and storage medium |
| CN116386848A (en) * | 2023-03-10 | 2023-07-04 | 王子骁 | Multidimensional thyroid nodule accurate evaluation system and method based on AI technology |
| CN116825336A (en) * | 2023-08-30 | 2023-09-29 | 山东志诚普惠健康科技有限公司 | AI-based medical information intelligent management method and system |
| CN117153419A (en) * | 2023-10-31 | 2023-12-01 | 湖北福鑫科创信息技术有限公司 | Data integration tool for medical institutions |
| CN117542467A (en) * | 2024-01-09 | 2024-02-09 | 四川互慧软件有限公司 | Automatic construction method of disease-specific standard database based on patient data |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101599088A (en) * | 2008-11-18 | 2009-12-09 | 北京美智医疗科技有限公司 | The mining multi-dimensional data system and method for medical information system |
| CN107680676A (en) * | 2017-09-26 | 2018-02-09 | 电子科技大学 | A kind of gestational diabetes Forecasting Methodology based on electronic health record data-driven |
| CN107908621A (en) * | 2017-11-16 | 2018-04-13 | 东华大学 | Tumor of breast risk assessment system based on ultrasonic examination report text data |
| CN108615560A (en) * | 2018-03-19 | 2018-10-02 | 安徽锐欧赛智能科技有限公司 | A kind of clinical medical data analysis method based on data mining |
| CN109243616A (en) * | 2018-06-29 | 2019-01-18 | 东华大学 | Mammary gland electronic health record joint Relation extraction and architectural system based on deep learning |
-
2019
- 2019-02-01 CN CN201910101985.XA patent/CN109830303A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101599088A (en) * | 2008-11-18 | 2009-12-09 | 北京美智医疗科技有限公司 | The mining multi-dimensional data system and method for medical information system |
| CN107680676A (en) * | 2017-09-26 | 2018-02-09 | 电子科技大学 | A kind of gestational diabetes Forecasting Methodology based on electronic health record data-driven |
| CN107908621A (en) * | 2017-11-16 | 2018-04-13 | 东华大学 | Tumor of breast risk assessment system based on ultrasonic examination report text data |
| CN108615560A (en) * | 2018-03-19 | 2018-10-02 | 安徽锐欧赛智能科技有限公司 | A kind of clinical medical data analysis method based on data mining |
| CN109243616A (en) * | 2018-06-29 | 2019-01-18 | 东华大学 | Mammary gland electronic health record joint Relation extraction and architectural system based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| 《信息系统审计研究报告》课题组: "《信息系统审计研究报告》", 30 November 2015, 中国时代经济出版社 * |
| 朱立峰等: "多中心临床大数据平台建设及深度应用", 《大数据》 * |
Cited By (109)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI723868B (en) * | 2019-06-26 | 2021-04-01 | 義守大學 | Method for applying a label made after sampling to neural network training model |
| CN110277172A (en) * | 2019-06-27 | 2019-09-24 | 齐鲁工业大学 | A kind of clinical application behavior analysis system and its working method based on efficient negative sequence mining mode |
| CN110477954A (en) * | 2019-07-08 | 2019-11-22 | 无锡海斯凯尔医学技术有限公司 | Detection device based on elastogram |
| CN110477954B (en) * | 2019-07-08 | 2021-07-27 | 无锡海斯凯尔医学技术有限公司 | Detection equipment based on elasticity formation of image |
| WO2020233254A1 (en) * | 2019-07-12 | 2020-11-26 | 之江实验室 | Medical data analysis system integrating structured image data |
| CN110415831A (en) * | 2019-07-18 | 2019-11-05 | 天宜(天津)信息科技有限公司 | A kind of medical treatment big data cloud service analysis platform |
| CN110415831B (en) * | 2019-07-18 | 2023-04-18 | 天宜(天津)信息科技有限公司 | Medical big data cloud service analysis platform |
| CN110570921A (en) * | 2019-08-20 | 2019-12-13 | 广东省第二中医院（广东省中医药工程技术研究院） | Clinical information processing system based on single disease category |
| CN110544528B (en) * | 2019-08-29 | 2022-06-07 | 中南大学 | Advanced learning-based upper and lower ophthalmic remote diagnosis platform and construction method thereof |
| CN110544528A (en) * | 2019-08-29 | 2019-12-06 | 中南大学 | advanced learning-based upper and lower ophthalmic remote diagnosis platform and construction method thereof |
| CN110767317A (en) * | 2019-08-30 | 2020-02-07 | 贵州力创科技发展有限公司 | Cloud computing platform and method based on data mining and big data analysis |
| CN110675952A (en) * | 2019-09-19 | 2020-01-10 | 上海腾程医学科技信息有限公司 | Checking decision method and device, terminal equipment and computer readable storage medium |
| CN110569372A (en) * | 2019-09-20 | 2019-12-13 | 四川大学 | construction method of heart disease big data knowledge graph system |
| CN110875095A (en) * | 2019-09-27 | 2020-03-10 | 长沙瀚云信息科技有限公司 | Standardized clinical big data center system |
| CN110737731B (en) * | 2019-10-25 | 2023-12-29 | 徐州工程学院 | Decision tree-based public accumulation user data refinement analysis system and method |
| CN110737731A (en) * | 2019-10-25 | 2020-01-31 | 徐州工程学院 | accumulation fund user data refinement analysis system and method based on decision tree |
| CN111081016A (en) * | 2019-12-18 | 2020-04-28 | 北京航空航天大学 | Urban traffic abnormity identification method based on complex network theory |
| CN111081016B (en) * | 2019-12-18 | 2021-07-06 | 北京航空航天大学 | Urban traffic abnormity identification method based on complex network theory |
| CN113053479A (en) * | 2019-12-27 | 2021-06-29 | 天津幸福生命科技有限公司 | Medical data processing method, device, medium and electronic equipment |
| CN111341456A (en) * | 2020-02-21 | 2020-06-26 | 中南大学湘雅医院 | Method and device for generating diabetic foot knowledge map and readable storage medium |
| CN111341456B (en) * | 2020-02-21 | 2024-02-23 | 中南大学湘雅医院 | Method and device for generating diabetic foot knowledge graph and readable storage medium |
| CN111382151A (en) * | 2020-03-23 | 2020-07-07 | 新疆大学 | CT medical image cleaning method based on data desensitization |
| CN111428930A (en) * | 2020-03-24 | 2020-07-17 | 中电药明数据科技(成都)有限公司 | GBDT-based medicine patient using number prediction method and system |
| CN111696675B (en) * | 2020-05-22 | 2023-09-19 | 深圳赛安特技术服务有限公司 | User data classification method and device based on Internet of things data and computer equipment |
| CN111696675A (en) * | 2020-05-22 | 2020-09-22 | 平安国际智慧城市科技股份有限公司 | User data classification method and device based on Internet of things data and computer equipment |
| CN111768846A (en) * | 2020-05-27 | 2020-10-13 | 医利捷(上海)信息科技有限公司 | Clinical data management method |
| CN111696665A (en) * | 2020-06-10 | 2020-09-22 | 杭州联众医疗科技股份有限公司 | Auxiliary decision making system based on time coding |
| CN111696666A (en) * | 2020-06-10 | 2020-09-22 | 杭州联众医疗科技股份有限公司 | Intelligent chronic disease management system based on time coding |
| CN111724898A (en) * | 2020-06-15 | 2020-09-29 | 中国医学科学院医学信息研究所 | Intelligent skin disease monitoring and early warning system based on big data technology |
| CN111696669A (en) * | 2020-06-15 | 2020-09-22 | 山东搜搜中医信息科技有限公司 | Quantitative diagnosis of clinical multidimensional data of traditional Chinese medicine |
| CN111710427A (en) * | 2020-06-17 | 2020-09-25 | 广州市金域转化医学研究院有限公司 | Cervical precancerous early lesion stage diagnosis model and establishment method |
| CN112150209A (en) * | 2020-06-19 | 2020-12-29 | 南京理工大学 | Construction method of CNN-LSTM time sequence prediction model based on clustering center |
| CN112150209B (en) * | 2020-06-19 | 2022-10-18 | 南京理工大学 | Construction method of CNN-LSTM time sequence prediction model based on clustering center |
| CN111739639A (en) * | 2020-06-19 | 2020-10-02 | 杭州联众医疗科技股份有限公司 | Perioperative complication data prediction system based on time coding |
| CN111863267B (en) * | 2020-07-08 | 2024-01-26 | 首都医科大学附属北京天坛医院 | Data information acquisition method, data analysis method, device and storage medium |
| CN111863267A (en) * | 2020-07-08 | 2020-10-30 | 首都医科大学附属北京天坛医院 | Data information acquisition method, data analysis device and storage medium |
| CN111949801A (en) * | 2020-07-27 | 2020-11-17 | 西北工业大学 | Knowledge graph fusion method of doctor experience knowledge and ultrasonic image information |
| CN111949801B (en) * | 2020-07-27 | 2023-10-24 | 西北工业大学 | Knowledge graph fusion method for doctor experience knowledge and ultrasonic image information |
| CN112037925B (en) * | 2020-07-29 | 2023-06-23 | 郑州大学第一附属医院 | LSTM algorithm-based early warning method for new major infectious diseases |
| CN112037925A (en) * | 2020-07-29 | 2020-12-04 | 郑州大学第一附属医院 | LSTM algorithm-based early warning method for newly-released major infectious diseases |
| CN111914026A (en) * | 2020-07-31 | 2020-11-10 | 南京朗赢信息技术有限公司 | General data exchange sharing service platform |
| CN111899828A (en) * | 2020-07-31 | 2020-11-06 | 青岛百洋智能科技股份有限公司 | Knowledge graph driven breast cancer diagnosis and treatment scheme recommendation system |
| CN112100286A (en) * | 2020-08-14 | 2020-12-18 | 华南理工大学 | Computer-aided decision-making method, device and system based on multi-dimensional data and server |
| CN111951976A (en) * | 2020-08-21 | 2020-11-17 | 上海交通大学医学院附属第九人民医院 | Value judgment method, system, terminal and medium based on medical data margin |
| CN111951976B (en) * | 2020-08-21 | 2024-03-22 | 上海交通大学医学院附属第九人民医院 | Value judging method, system, terminal and medium based on medical data allowance |
| CN111984987B (en) * | 2020-09-01 | 2024-04-02 | 上海梅斯医药科技有限公司 | Method, device, system and medium for desensitizing and restoring electronic medical records |
| CN111984987A (en) * | 2020-09-01 | 2020-11-24 | 上海梅斯医药科技有限公司 | Method, device, system and medium for desensitization and reduction of electronic medical record |
| CN112286985A (en) * | 2020-10-13 | 2021-01-29 | 江苏云脑数据科技有限公司 | Clinical research statistical analysis system based on cloud computing |
| CN112200374A (en) * | 2020-10-15 | 2021-01-08 | 平安国际智慧城市科技股份有限公司 | Medical data processing method, device, electronic equipment and medium |
| CN112185585A (en) * | 2020-11-03 | 2021-01-05 | 浙江大学滨海产业技术研究院 | Diabetes early warning method based on metabonomics |
| CN112380763A (en) * | 2020-11-03 | 2021-02-19 | 浙大城市学院 | System and method for analyzing reliability of in-pile component based on data mining |
| WO2021175038A1 (en) * | 2020-11-13 | 2021-09-10 | 之江实验室 | Patient data visualization method and system for assisting decision-making in chronic disease |
| US11521751B2 (en) * | 2020-11-13 | 2022-12-06 | Zhejiang Lab | Patient data visualization method and system for assisting decision making in chronic diseases |
| CN112365976A (en) * | 2020-11-14 | 2021-02-12 | 南昌大学第二附属医院 | Compound disease clinical path construction method and system based on transfer learning |
| CN112365976B (en) * | 2020-11-14 | 2023-08-11 | 南昌大学第二附属医院 | Composite disease species clinical path construction method and system based on transfer learning |
| CN112446862A (en) * | 2020-11-25 | 2021-03-05 | 北京医准智能科技有限公司 | Dynamic breast ultrasound video full-focus real-time detection and segmentation device and system based on artificial intelligence and image processing method |
| CN112635074A (en) * | 2020-12-21 | 2021-04-09 | 云南省疾病预防控制中心 | AIDS prevention and treatment decision method based on multi-data analysis model |
| CN112786126B (en) * | 2020-12-31 | 2023-11-03 | 天津开心生活科技有限公司 | Time sequence analysis method and device for clinical test data, electronic equipment and medium |
| CN112786126A (en) * | 2020-12-31 | 2021-05-11 | 天津开心生活科技有限公司 | Time sequence analysis method and device of clinical test data, electronic equipment and medium |
| CN112988783A (en) * | 2021-03-12 | 2021-06-18 | 李涛 | Public opinion occurrence time sequence analysis method based on multidimensional data model |
| CN112927810A (en) * | 2021-03-23 | 2021-06-08 | 崔剑虹 | Smart medical response method based on big data and smart medical cloud computing system |
| CN113539471A (en) * | 2021-03-26 | 2021-10-22 | 内蒙古卫数数据科技有限公司 | Auxiliary diagnosis method and system for hyperplasia of mammary glands based on conventional inspection data |
| CN113096817A (en) * | 2021-04-13 | 2021-07-09 | 北京大学 | Method, apparatus, computer device and storage medium for disease prediction |
| CN113160999A (en) * | 2021-04-25 | 2021-07-23 | 厦门拜特信息科技有限公司 | Data structured analysis system and data processing method for medical decision |
| WO2022228473A1 (en) * | 2021-04-27 | 2022-11-03 | 联峰远程健康管理服务有限公司 | Smart health management system for use in telemedicine service and method used in same |
| CN113177040A (en) * | 2021-04-29 | 2021-07-27 | 东北大学 | Full-process big data cleaning and analyzing method for aluminum/copper plate strip production |
| CN113380360A (en) * | 2021-06-07 | 2021-09-10 | 厦门大学 | Similar medical record retrieval method and system based on multi-mode medical record map |
| CN113314201A (en) * | 2021-06-17 | 2021-08-27 | 南通市第一人民医院 | Neurology clinical nursing potential safety hazard analysis method and system |
| CN113314201B (en) * | 2021-06-17 | 2022-05-13 | 南通市第一人民医院 | Potential safety hazard analysis method and system for neurology clinical nursing |
| CN113436745A (en) * | 2021-06-30 | 2021-09-24 | 四川大学华西医院 | Artificial intelligence auxiliary diagnosis method based on database analysis |
| CN113609555A (en) * | 2021-07-16 | 2021-11-05 | 黄河勘测规划设计研究院有限公司 | Hydraulic metal structure design method based on big data technology |
| CN113609555B (en) * | 2021-07-16 | 2023-10-20 | 黄河勘测规划设计研究院有限公司 | Hydraulic metal structure design method based on big data technology |
| CN113436747A (en) * | 2021-07-20 | 2021-09-24 | 四川省医学科学院·四川省人民医院 | Medical data clinical auxiliary system and method based on medical data analysis model |
| CN113674867A (en) * | 2021-07-27 | 2021-11-19 | 上海药慧信息技术有限公司 | Clinical data mining method and device, electronic equipment and storage medium |
| CN113539414A (en) * | 2021-07-30 | 2021-10-22 | 中电药明数据科技(成都)有限公司 | Method and system for predicting rationality of antibiotic medication |
| CN113628734A (en) * | 2021-08-02 | 2021-11-09 | 浙江海心智惠科技有限公司 | Design method of oncology electronic medical advice system with clinical decision intelligent recommendation function |
| CN113688169A (en) * | 2021-08-11 | 2021-11-23 | 北京科技大学 | Mine potential safety hazard identification and early warning system based on big data analysis |
| CN113688169B (en) * | 2021-08-11 | 2023-08-08 | 北京科技大学 | Mine potential safety hazard identification and early warning system based on big data analysis |
| CN113823421B (en) * | 2021-08-20 | 2024-02-13 | 武汉心络科技有限公司 | Information providing method, apparatus, device and storage medium |
| CN113823421A (en) * | 2021-08-20 | 2021-12-21 | 武汉心络科技有限公司 | Information providing method, device, equipment and storage medium |
| CN113889279A (en) * | 2021-09-28 | 2022-01-04 | 北京华彬立成科技有限公司 | Combination therapy information mining and inquiring method, device and electronic equipment |
| CN113611411B (en) * | 2021-10-09 | 2021-12-31 | 浙江大学 | Body examination aid decision-making system based on false negative sample identification |
| WO2023056918A1 (en) * | 2021-10-09 | 2023-04-13 | 浙江大学 | False negative sample recognition-based physical examination assistant decision-making system |
| CN113611411A (en) * | 2021-10-09 | 2021-11-05 | 浙江大学 | Body examination aid decision-making system based on false negative sample identification |
| CN113744873B (en) * | 2021-11-08 | 2022-02-11 | 浙江大学 | Heating to-be-checked auxiliary differential diagnosis system based on task decomposition strategy |
| CN113744873A (en) * | 2021-11-08 | 2021-12-03 | 浙江大学 | Heating to-be-checked auxiliary differential diagnosis system based on task decomposition strategy |
| CN114121293A (en) * | 2021-11-12 | 2022-03-01 | 北京华彬立成科技有限公司 | Clinical trial information mining and inquiring method and device |
| CN114068013B (en) * | 2021-11-16 | 2022-09-23 | 高峰 | Cerebral artery occlusion artificial intelligence assistant decision system |
| CN114068013A (en) * | 2021-11-16 | 2022-02-18 | 高峰 | Cerebral artery occlusion artificial intelligence assistant decision system |
| CN114496177A (en) * | 2022-01-24 | 2022-05-13 | 佳木斯大学 | Method and system for detecting clinical infection source of infectious department based on big data |
| CN114464314A (en) * | 2022-02-08 | 2022-05-10 | 四川大学华西医院 | Clinical body symptom classification diagnosis system |
| CN114678132A (en) * | 2022-02-22 | 2022-06-28 | 北京颐圣智能科技有限公司 | Self-learning medical wind control system and method based on clinical behavior feedback |
| CN114678132B (en) * | 2022-02-22 | 2023-07-18 | 北京颐圣智能科技有限公司 | Self-learning medical wind control system and method based on clinical behavior feedback |
| CN114564755A (en) * | 2022-03-03 | 2022-05-31 | 曾迎春 | Cancer data management platform based on block chain technology |
| CN114912804A (en) * | 2022-05-17 | 2022-08-16 | 四川大学华西医院 | Scientific research data related property control method and system |
| CN115145993A (en) * | 2022-07-05 | 2022-10-04 | 西南交通大学 | Railway freight big data visualization display platform based on self-learning rule operation |
| CN115083601A (en) * | 2022-07-25 | 2022-09-20 | 四川省医学科学院·四川省人民医院 | Type 2diabetes auxiliary decision making system based on machine learning |
| CN115083616A (en) * | 2022-08-16 | 2022-09-20 | 之江实验室 | Chronic nephropathy subtype mining system based on self-supervision graph clustering |
| JP7404581B1 (en) | 2022-08-16 | 2023-12-25 | 之江実験室 | Chronic nephropathy subtype mining system based on self-supervised graph clustering |
| CN115083616B (en) * | 2022-08-16 | 2022-11-08 | 之江实验室 | Chronic nephropathy subtype mining system based on self-supervision graph clustering |
| CN115240800A (en) * | 2022-09-26 | 2022-10-25 | 北京泽桥医疗科技股份有限公司 | Medical data intelligent analysis execution method based on big data platform |
| CN115617840B (en) * | 2022-12-19 | 2023-03-10 | 江西曼荼罗软件有限公司 | Medical data retrieval platform construction method, system, computer and storage medium |
| CN115617840A (en) * | 2022-12-19 | 2023-01-17 | 江西曼荼罗软件有限公司 | Medical data retrieval platform construction method, system, computer and storage medium |
| CN116386848A (en) * | 2023-03-10 | 2023-07-04 | 王子骁 | Multidimensional thyroid nodule accurate evaluation system and method based on AI technology |
| CN116825336A (en) * | 2023-08-30 | 2023-09-29 | 山东志诚普惠健康科技有限公司 | AI-based medical information intelligent management method and system |
| CN117153419A (en) * | 2023-10-31 | 2023-12-01 | 湖北福鑫科创信息技术有限公司 | Data integration tool for medical institutions |
| CN117153419B (en) * | 2023-10-31 | 2024-01-26 | 湖北福鑫科创信息技术有限公司 | Data integration tool for medical institutions |
| CN117542467A (en) * | 2024-01-09 | 2024-02-09 | 四川互慧软件有限公司 | Automatic construction method of disease-specific standard database based on patient data |
| CN117542467B (en) * | 2024-01-09 | 2024-04-12 | 四川互慧软件有限公司 | Automatic construction method of disease-specific standard database based on patient data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109830303A (en) | Clinical data mining analysis and aid decision-making method based on internet integration medical platform | |
| Ambekar et al. | Disease risk prediction by using convolutional neural network | |
| US10824607B2 (en) | Topological data analysis of data from a fact table and related dimension tables | |
| Milovic et al. | Prediction and decision making in health care using data mining | |
| ȚĂRANU | Data mining in healthcare: decision making and precision | |
| US20180025093A1 (en) | Query capabilities of topological data analysis graphs | |
| CN109785927A (en) | Clinical document structuring processing method based on internet integration medical platform | |
| CN109841282A (en) | A kind of Chinese medicine health control cloud system and its building method based on cloud computing | |
| Dey et al. | Study and analysis of data mining algorithms for healthcare decision support system | |
| Djatna et al. | An intuitionistic fuzzy diagnosis analytics for stroke disease | |
| CN108962394B (en) | Medical data decision support method and system | |
| Chauhan et al. | A robust model for big healthcare data analytics | |
| CN116864139A (en) | Disease risk assessment method, device, computer equipment and readable storage medium | |
| Chou et al. | Extracting drug utilization knowledge using self-organizing map and rough set theory | |
| Datta et al. | Development of predictive model of diabetic using supervised machine learning classification algorithm of ensemble voting | |
| Ahmad | Mining health data for breast cancer diagnosis using machine learning | |
| Li et al. | Study of E-business applications based on big data analysis in modern hospital health management | |
| Retal et al. | Machine learning for diabetes prediction: a systematic review and a conceptual framework for early prediction | |
| Kamble et al. | Smart Health Prediction System Using Data Mining | |
| Jaber et al. | Early prediction of diabetic using data mining | |
| Gu et al. | Which is more reliable, expert experience or information itself? weight scheme of complex cases for health management decision making | |
| Jia et al. | Dkdr: An approach of knowledge graph and deep reinforcement learning for disease diagnosis | |
| Darabi et al. | A Case-Based-Reasoning System for Feature Selection and Diagnosing Asthma | |
| Saadi et al. | Integration of fuzzy clustering into the case base reasoning for the prediction of response to immunotherapy treatment | |
| Shah et al. | A Review on Big Data Practices in Healthcare |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190531 |