CN109801687A - A kind of construction method and system of the causality knowledge base towards medicine - Google Patents
A kind of construction method and system of the causality knowledge base towards medicine Download PDFInfo
- Publication number
- CN109801687A CN109801687A CN201910034537.2A CN201910034537A CN109801687A CN 109801687 A CN109801687 A CN 109801687A CN 201910034537 A CN201910034537 A CN 201910034537A CN 109801687 A CN109801687 A CN 109801687A
- Authority
- CN
- China
- Prior art keywords
- illness
- relationship
- cause
- effect
- knowledge base
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/041—Abduction
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Pathology (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The construction method for the causality knowledge base towards medicine that the present invention relates to a kind of, its step include: literary unit can obtain numerous pertinent literatures containing various diseases and it is carried out classification form several literary unit bodies to construct original library, so that data cell can be obtained main characteristic parameters based on literary unit body and construct data set based on main characteristic parameters, cause and effect unit is based on main characteristic parameters and data set constructs Bayesian network, to analyze the average cause-effect between illness by data pattern, to which blocks of knowledge can construct knowledge base in a manner of forming the corresponding relationship of the average cause-effect between illness and between illness based on pertinent literature.
Description
Technical field
The present invention relates to medical information technical field more particularly to a kind of structures of the causality knowledge base towards medicine
Construction method and system.
Background technique
During scientific research, document is a kind of carrier for recording scientific achievement and most having convincingness and conviction power.
Beyond all doubt says, the most content of scientific research activity is had recorded in document.A large amount of document describes the pass between illness
Connection, but it is very difficult that doctor, which needs to consult the causality that a large amount of document is gone between research and illness,.Such as, in medicine
Boundary, complication and complication be after a kind of disease has occurred, follow this disease have occurred another or it is a variety of
Disease.Wherein, complication and complication are a complicated clinical medicine concepts.Complication refers to a kind of disease in development process
Cause the generation of another disease or symptom；And complication refers to during diagnosis and nursing, patient is by suffering from a kind of disease merging
Another or several diseases related with this disease have occurred.In medical field, medical research it is emphasised that causality without
It is incidence relation.And between complication and complication disease, have causality between complication and primary disease, and complication with
Noncausal relationship between principal disease.Therefore, whether it is causality that doctor can be found from numerous documents between illness
It is the technical issues that need to address.
For example, a kind of case history statistics of complication and complication disclosed in the Chinese patent of Publication No. CN107145712A
Analysis system.The system includes the Chi-square Test list of diagnostic code maintenance unit, case history counting unit, 2 × 2 cross division data
Member and report generation unit safeguard the mapping relations between each ID and diagnosis, and establish diagnosis ID diagnosis statistical form；Its
In, the file of diagnosis ID statistical form corresponds to other diagnosis in case history homepage discharge diagnosis；Hospital is imported by data-interface again
K parts of history case history, according to diagnosis ID statistical form, the Main Diagnosis in case history discharge diagnosis is converted into ID with other diagnosis and is examined
It is disconnected, it is then calculated by the Chi-square Test of 2 × 2 cross division data and Main Diagnosis chi-square value corresponding with various diagnosis,
And be ranked up, so, it is possible to analyze rapidly using computer high-speed computation may have causality with selected Main Diagnosis
Complication and complication, reach the better beneficial effect of speed.
For example, a kind of complication and complication impact factor disclosed in the Chinese patent of Publication No. CN107799182A
Evaluation method and electronic equipment.This method seeks impact factor vector by excellent calculation algorithm so that Diagnosis vector and influence because
The closest resource consumption level parameter of the inner product of subvector, so as to get each under a diagnosis associated packets
Complication and complication occur whether or severity for medical resource consumption level weighing factor.The weighing factor can
With accurately estimate each complication and complication whether appearance or severity for medical resource consumption level shadow
It rings.
For example, a kind of assessment chronic disease Cost Modeling method disclosed in the Chinese patent of Publication No. CN106407686A.
This method passes through screening sample and feature selecting first, then uses regression model, obtains each impact factor to chronic disease expense
Influence.The invention can directly quantify influence degree of the complication complication to its data expense of chronic disease out, be chronic
And medical treatment control take offer foundation.
For example, a kind of inpatient's medical control quality evaluation disclosed in the Chinese patent of Publication No. CN105046406A
Method, this method comprises: historical data screening and modeling；Data identify and cleaning；Medical diagnosis on disease associated packets DRG and model
Classification；ICD closes complication and its classification set of its dependent variable when being admitted to hospital；Be admitted to hospital and close the statistical check of complication variable with
Screening；The foundation of mathematical modeling and quality verification；Current data screening is calculated with pre-value；Calculate the risk profile of patient admission
Value, realize to each inpatient the death rate, length of stay and medical treatment cost risk profile of being admitted to hospital.The invention is by counting greatly
Effective conversion of the medical data from data to solution is realized according to the methods of analysis, mathematical statistics and machine learning, it is real
Data value is showed.It solves incomparable problem between medical data, can not only realize that the quality of medical care between disease is commented
Estimate, can also be achieved between doctor, commented between hospital department, between hospital in the performance reasonability of inpatient's disease treatment management
Valence.
For example, a kind of introducing radiation responsive parameter normal tissue disclosed in Publication No. CN102542153B is simultaneously
The method for sending out disease impact probability.This method is based on simple normal tissue organ model, and dosage and the life of organ model is calculated
Object effective dose BED distribution returns case processing BED distribution that total existence score SF is calculated, organ model is calculated using SF
50% complication of the NTCPLKB model effective uniform dose EUD of corresponding broad sense is utilized EUD model by effective uniform dose EUD
In include a variety of radiosusceptibility parameters, introduce radiation sensitivity parameters on normal tissue complication probability influence.
For example, one kind disclosed in the Chinese patent of Publication No. CN106295187A is supported towards intelligent clinical aid decision
The construction of knowledge base method and system of system, this method include obtain input information, to the input information progress word segmentation processing,
Part-of-speech tagging and syntactic analysis obtain relational dependence tree, extract concept in the relational dependence tree, entity, entity and modify
Language；And according to the concept, the entity, the entity modifier, by relationship semantical definition rule, obtain the relationship according to
Relationship between each entity of Lai Shuzhong；Setting extension triple, by the extension triple by the relational dependence tree
In relationship between each entity stored, to complete building knowledge base.The invention can be used for Clinic Case mostly and feature
More situations can realize flexible expansion for case history representation manners.But this application only passes through the short of recognition expression semantic relation
Language extracts the relationship between entity, and relationship lays particular emphasis on correlativity, and correlativity not necessarily causality, therefore
Causality acquired in the patent is unreliable.
For example, a kind of congenital cataract art time complication disclosed in the Chinese patent of Publication No. CN106667443A is pre-
Survey method and system.This method obtains predictive factor by clinical information；By predictive factor by NB Algorithm, obtain
Prediction result；Prediction result is presented；Corresponding follow-up information is obtained according to prediction result, so as to Accurate Prediction complication
Occur.
By the above prior art investigation it can be found that for whether constituting showing for complication and complication between illness
Having that technology is much insufficient, only Publication No. CN107145712A is only capable of being determined by simple mathematical statistics, and
And determined with independence assumption, it is merely capable of illustrating the relevance between illness, can not illustrate the cause and effect between illness
Property, this is not enough to quantify the causality between illness.
Summary of the invention
For the deficiencies of the prior art, the building side for the causality knowledge base towards medicine that the present invention provides a kind of
Method is related to whether constituting the building side of the knowledge base of complication or complication between a kind of various diseases based on literature search
Method, step include: literary unit building original library；Data cell constructs data set；Between cause and effect unit illness because
Fruit relationship；Blocks of knowledge stores the original library, the data set and/or the average cause-effect and can be read with constructing
The knowledge base for taking and/or showing is supplied to medical worker reference, study and/or decision in the form of data quantization；Institute
State literary unit can obtain numerous pertinent literatures containing various diseases and it is carried out classification form several literary unit bodies
To construct original library, so that the data cell can obtain main characteristic parameters and base based on the literary unit body
Data set is constructed in the main characteristic parameters, the cause and effect unit is based on the main characteristic parameters and the data set constructs
Bayesian network, to analyze the average cause-effect between illness by data pattern, so that the blocks of knowledge can be based on
The pertinent literature is to form the side of the corresponding relationship of the average cause-effect between the illness and between the illness
Formula constructs the knowledge base.The present invention is to be not related to disease by the way that the initial data recorded in document is formed a kind of knowledge base
The diagnostic and therapeutic method of disease.
According to a kind of preferred embodiment, the pertinent literature classification is carried out as follows: the literary unit
The frequency for counting words/phrases in each document obtains the word/phrase joint according to independence assumption and occurs
Probability；The literary unit calculates the word/phrase relevance intensity, and corrects institute based on the relevance intensity
Joint probability of occurrence is stated to obtain the association reduced coordinate of the document；The association reduction that the literary unit constructs the document is sat
Mark, and the association reduced coordinate based on whole pertinent literatures and the classification function of relevance intensity building will
The pertinent literature is classified according to the form of iterative algorithm to form several literary unit bodies；Wherein, the classification
Function can carry out deep learning based on the sample size of the pertinent literature, to enhance the precision of the literary unit.
According to a kind of preferred embodiment, in the case where the data cell obtains the literary unit body, institute
It states data cell and obtains the data set in such a way that illness is to pairing；The data cell is each pertinent literature
The relationship between wherein illness pair is extracted in a manner of the syntactic analysis of natural language processing, to establish the relationship of the illness pair
Knowledge base, the relationship between the illness pair includes positive relationship, inverse relationship and vertical relation；Also, the data cell
The document for including the illness pair is retrieved in a manner of fusion in the literary unit body based on relationship knowledge table
Obtain relationship certainty value library of the relationship certainty value of the illness pair to establish the illness pair, the pass between the illness pair
System includes positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value；To which the data cell is based on to institute
There are the relational knowledge base established in the way of matching two-by-two between illness and relationship certainty value library to construct the data set.
According to a kind of preferred embodiment, the cause and effect unit 3 constructs Bayesian network as follows, S31: base
Bayesian network evaluation function is constructed in the relational knowledge base:
logP(G,D,KL)=logP (G)+logP (D | G)+logP (KL|G)
S32: non-directed graph structural constraint is constructed based on the relational knowledge base；For the data set D, to the data set
Any illness in D is to LmAnd Ln, attribute is obtained to L to the relational knowledge base by retrieval illnessmAnd LnIllness to volume
Number, according to the illness to illness in number searching document to LmAnd LnL in relationship certainty value tablem→LnRelationship certainty value
And Ln→LmRelationship certainty value,
S33: Bayesian network is constructed based on the Bayesian network evaluation function and the non-directed graph structural constraint.
According to a kind of preferred embodiment, the cause and effect unit is based on the Bayesian network and Pearl principle calculates
Average cause-effect between each illness pair, when the average cause-effect is more than the cause-effect threshold value of setting, illness it
Between constitute complication；When the average cause-effect is no more than the cause-effect threshold value of setting, complication is constituted between illness.
According to a kind of preferred embodiment, for illness Lm, by way of traversal about based on the undirected graph structure
Beam obtains and illness LmThe node being connected constitutes its node collection；And gradually calculate each node and illness LmBetween correlation,
And therefrom choose correlation maximum node carry out independence assumption, delete at data-oriented collection D with LmIndependent node；
Illness LnWith illness LmBetween independence pass through mutual information measure:
When the mutual information has exceeded the threshold value of mutual information, then illness LnWith illness LmIt is less independent with correlation；
In the threshold value of the mutual information without departing from mutual information, then illness LnWith illness LmIt is independent without correlation.
According to a kind of preferred embodiment, the structure for the causality knowledge base towards medicine that the invention also discloses a kind of
System is built, which includes literary unit: for constructing original library；Data cell: for constructing data set；Cause and effect list
Member: for the average cause-effect between illness；And blocks of knowledge: for store the original library, the data set and/
Or the average cause-effect is supplied in the form of data quantization with constructing the knowledge base that can be read and/or show
Medical worker reference, study and/or decision；The literary unit can be numerous containing more based on user-defined request
It plants the pertinent literature of illness and carries out classification to it and form several literary unit bodies to construct original library, so that the number
Main characteristic parameters can be obtained based on the literary unit body according to unit and construct data set based on the main characteristic parameters,
To reduce numerous characteristic parameters of numerous pertinent literatures formation for interference causal between illness and improve original
The utility value of document databse；The cause and effect unit is based on the main characteristic parameters and the data set constructs Bayesian network,
To excavate the average cause-effect between illness by data pattern, so as to according between the average cause-effect illness
Whether complication or complication are constituted.
According to a kind of preferred embodiment, the literary unit counts the frequency of words/phrases in each document, presses
The word/phrase joint probability of occurrence is obtained according to independence assumption；The literary unit calculating word/described
The relevance intensity of phrase, and the joint probability of occurrence is corrected to obtain the association of the document about based on the relevance intensity
Change coordinate；The literary unit constructs the association reduced coordinate of the document, and based on described in whole pertinent literatures
Be associated with reduced coordinate and the relevance intensity building classification function by the pertinent literature according to iterative algorithm form into
Row classification is to form several literary unit bodies；Wherein, the classification function can be based on the sample size of the pertinent literature
Deep learning is carried out, to enhance the precision of the literary unit.
According to a kind of preferred embodiment, in the case where the data cell obtains the literary unit body, institute
State data cell and obtain the data set in such a way that illness is to pairing: the data cell is each pertinent literature
The relationship between wherein illness pair is extracted in a manner of the syntactic analysis of natural language processing, to establish the relationship of the illness pair
Knowledge base, the relationship between the illness pair includes positive relationship, inverse relationship and vertical relation；Also, the data cell
The document for including the illness pair is retrieved in a manner of fusion in the literary unit body based on relationship knowledge table
Obtain relationship certainty value library of the relationship certainty value of the illness pair to establish the illness pair, the pass between the illness pair
System includes positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value；To which the data cell is based on to institute
There are the relational knowledge base established in the way of matching two-by-two between illness and relationship certainty value library to construct the data set.
According to a kind of preferred embodiment, the cause and effect unit constructs Bayesian network as follows, S31: base
Bayesian network evaluation function is constructed in the relational knowledge base:
logP(G,D,KL)=logP (G)+log (D | G)+logP (KL|G)
S32: non-directed graph structural constraint is constructed based on the relational knowledge base；To the data set D, to the data set D
In any illness to LmAnd Ln, attribute is obtained to L to the relational knowledge base by retrieval illnessmAnd LnIllness to number,
According to the illness to illness in number searching document to LmAnd LnL in relationship certainty value tablem→LnRelationship certainty value and Ln
→LmRelationship certainty value,
S33: Bayesian network is constructed based on the Bayesian network evaluation function and the non-directed graph structural constraint.
The present invention provides a kind of building system of causality knowledge base towards medicine, and the present invention is based on existing concurrent
Disease/complication field documents and materials, construct original library, and design Bayesian network evaluation function, and construct pattra leaves
Then the undirected constraint diagram of this network constructs Bayesian network, the causality analysis reasoning between illness is realized, to analyze illness
Between causality.
Detailed description of the invention
Fig. 1 is a kind of preferred flow diagram of construction method provided by the invention；With
Fig. 2 is a kind of preferred module diagram of building system provided by the invention.
Reference signs list
1: literary unit 2: data cell
3: cause and effect unit 4: blocks of knowledge
Specific embodiment
It 1 and 2 is described in detail with reference to the accompanying drawing.
It in description of the invention, term " first ", " second ", " third " and is used for description purposes only, and should not be understood as
Indication or suggestion relative importance or the quantity for implicitly indicating indicated technical characteristic." first ", " are defined as a result,
Two ", " third " and feature can explicitly or implicitly include one or more of the features.In description of the invention
In, the meaning of " plurality " is two or more, unless otherwise specifically defined.
Embodiment 1
During scientific research, document is a kind of carrier for recording scientific achievement and most having convincingness and conviction power.
Beyond all doubt says, the most content of scientific research activity is had recorded in document.A large amount of document describes the pass between illness
Connection, but it is very difficult that doctor, which needs to consult the causality that a large amount of document is gone between research and illness,.For example, curing
Educational circles, there is no causalities between the illness of complication, and have causality between the illness of complication.Based on this this hair
It is bright by be collected, separate containing the document between various diseases, the processes such as parameter extraction and judgement, to pass through quantization
Mode provides a kind of construction method of the knowledge base of complication or complication, can pass through for doctor in the decision for the treatment of means
The mode of " data are spoken " provides strong reference.In the present invention, it is only in the document for polydispersion of comforming, Research Literature
The causality between illness referred to, with a kind of this knowledge base of being formed based on retrieval.The building system is not a kind of disease
Disease diagnoses and/or treats method.
The construction method for the causality knowledge base towards medicine that present embodiment discloses a kind of, do not cause conflict or
In contradictory situation, the entirety and/or partial content of the preferred embodiment of other embodiments can be used as the benefit of the present embodiment
It fills.Preferably, this method can be realized by method of the invention and/or other alternative modules.
Whether the building side of the knowledge base of complication or complication is constituted between a kind of various diseases based on literature search
Method, as shown in Figure 1, its step includes:
S1: literary unit 1 constructs original library；
S2: data cell 2 constructs data set；
S3: the causality between 3 illness of cause and effect unit；
S4: blocks of knowledge 4 stores the original library, the data set and/or the average cause-effect to construct
The knowledge base that can be read and/or show.To which the information that knowledge base 4 provides can be supplied in the form of data quantization
Medical worker reference, study and/or decision.
It is done in order to reduce numerous characteristic parameters that numerous pertinent literatures are formed for causal between illness pair
Disturb and improve the utility value in original library.Preferably, literary unit can obtain numerous correlation texts containing various diseases
It offers and carries out classification to it and form several literary unit bodies to construct original library, so that data cell can be based on document
Cell cube obtains main characteristic parameters and constructs data set based on main characteristic parameters.
Preferably, cause and effect unit is based on main characteristic parameters and data set constructs Bayesian network, to pass through data pattern
Analyze illness between average cause-effect, thus blocks of knowledge can based on pertinent literature be formed between illness and illness it
Between the mode of corresponding relationship of average cause-effect construct knowledge base.For example, the average cause-effect between illness can be anti-
Whether mirror between illness is to constitute complication and complication.
Preferably, numerous pertinent literatures containing various diseases of the literary unit 1 based on acquisition.1 pair of correlation of literary unit
Document carries out classification and forms several literary unit bodies to construct original library.The pertinent literature includes medical case history, research report
Announcement, meeting paper, periodical literature, books, academic paper and patent.In the case where such a large amount of document, need according to
Certain method is classified.Document classification is carried out to be to be able to the association effectively observed between illness and reduce system
Load.Such as it can classify according to disease of digestive tract, cardiovascular disease and neuropathy etc..It can also be according to science
Field is classified, such as rehabilitation and psychology etc. are classified.But, quasi- under the serious stern form of lot of documents
Really efficient classification will have a direct impact on the difference of complication and complication.Preferably, document point come can using Bayesian Method,
SVM method and k-NN method.
Preferably, pertinent literature classification is carried out as follows: S11: literary unit 1 counts single in each document
Word/phrase frequency obtains the joint probability of occurrence of words/phrases according to independence assumption.For example, for a specific text
It offers, joint probability of occurrence distribution can be calculated according to Nae Bayesianmethod.
S12: the relevance intensity of the calculating words/phrases of literary unit 1.By the calculating of relevance intensity, it is able to reflect
The relevance of words/phrases, the classification for document are suitable.Preferably, in classification, the collection that N is document sample is defined
It closes, V is the set of document type, ViIt is the subset of i-th of document type.W is words/phrases set, WiIt is i-th of word/word
The subset of group.In ViIn contain SjA sample, wherein the association reduced coordinate T of p-th of samplepIt is a n dimension group:
Wherein, ki(i=1,2,3 ... n) in i-th of word occur number,Normalization coefficient.
In ViInterconnection vector be all ViMiddle sample association reduced coordinate is averaged, which reflects word/word in document
Group relevance intensity i.e.:
S13: literary unit 1 obtains the association reduced coordinate of document, and is sat based on the association reduction of whole pertinent literatures
The classification function of mark building classifies pertinent literature according to the form of iterative algorithm to form several literary unit bodies.It is preferred that
Ground, for any document, it is associated with reduced coordinate are as follows:
In formula, qiIt is the number that i-th of word occurs in document.It is document to be sorted and every a kind of text carrying out classification
Offer ViSupporting point (b1,b2,…,bn) distance be denoted as:
According to relevance intensity, document classification function is constructed:
In formula, γiIt is related to relevance intensity.
Preferably, iterative algorithm can be using minimum iterative algorithm, minimum Optimized Iterative algorithm and expectation greatest iteration
Algorithm.Preferably, classification function can carry out deep learning based on the sample size of pertinent literature, to enhance literary unit 1
Precision.
Preferably, data cell 2 can be obtained main characteristic parameters based on literary unit body and be based on main characteristic parameters
Construct data set.To reduce numerous characteristic parameters of numerous pertinent literatures formation for interference causal between illness and mention
The utility value of high original document databse.Preferably, in the case where data cell 2 obtains literary unit body, data cell 2 is pressed
Data set is obtained according to mode of the illness to pairing.Data cell 2 is to each pertinent literature with the syntax of natural language processing point
Analysis mode extracts the relationship between wherein illness pair, the relationship packet to establish the relational knowledge base of illness pair, between illness pair
Include positive relationship, inverse relationship and vertical relation.Also, data cell 2 is based on relationship knowledge table in literary unit body to packet
Document containing illness pair is retrieved pass of the relationship certainty value that illness pair is obtained in a manner of fusion to establish illness pair
It is certainty value library, the relationship between illness pair includes positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value.
To which data cell 2 is based on to the relational knowledge base and relationship reliability established in the way of matching two-by-two between all illnesss
It is worth library and constructs data set.For example, obtaining illness L in pertinent literature1With illness L2.Illness L1With illness L2The relationship of appearance
It may be positive relationship, i.e. illness L1Influence illness L2, it is denoted as L1→L2.Illness L1With illness L2The relationship that can occur is possible
For inverse relationship, i.e. illness L2Influence illness L1, it is denoted as L2→L1.Illness L1With illness L2The relationship of appearance may close to be vertical
System, i.e. illness L2With illness L1Be independent of each other L1⊥L2.It can also include illness L since complication or complication are a variety of3
With illness L4Etc. several illnesss.According to the relationship of the above building illness, illness L can be constructed1With illness L3Relationship knowledge
Library, illness L2With illness L3Relational knowledge base, and so on.Then in unit document body, according to the content in different documents
Based on above-mentioned relation construction of knowledge base relationship certainty value library.Preferably, positive relationship certainty value, inverse relationship certainty value and vertical
The sum of relationship certainty value three is according to normalized.I.e. in unit document body, traversal queries are carried out to all documents, it is right
Positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value assign weight according to frequency.Data cell 2 will be above-mentioned
In relational knowledge base and relationship certainty value library building data set input cause and effect unit 3, carry out in next step.
Preferably for periodical literature, L1→L2Positive relationship certainty value can also be according to such as giving a definition:
Wherein, C (Xi) is the confidence level of document Xi, formula are as follows: C (Xi)=(IFi+1) × (CIi+1), Xi indicate i-th
Document, IFi are the impact factor after the standardization of periodical where document Xi, and CIi is the reference amount after standardization.If without document
There are L1And L2Relationship, then KL (L1→L2)=0, KL (L2→L1)=0, KL (L1⊥L2)=1.Other kinds of document can
To define in the same way, such as case history can be defined according to the authority of doctor.For meeting article, Ke Yigen
Be defined according to the authority of meeting etc..
Preferably, cause and effect unit 3 is based on main characteristic parameters and data set constructs Bayesian network.Preferably, main special
Levying parameter includes positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value.The cause and effect unit 3 is according to such as
Under type constructs Bayesian network:
S31: preferably, data set D=(D is defined1, D2……Di) it is several groups illness, L=(L1, L2……Ln) a certain
The specific set of disorders of group illness.Bayesian network evaluation function is constructed based on the relational knowledge base:
logP(G,D,KL)=logP (G)+logP (D | G)+logP (KL|G)
In formula, G is Bayes's grid, and value includes with L=(L1, L2……Ln) a certain group of illness specific illness
Collection is combined into the directed acyclic graph of node.Wherein, P (G) is prior distribution.According to existing knowledge it is found that logP (G)+logP (D | G)
Maximum value be equivalent to logP (G | D).LogP (G | D) it can be scored according to Bayesian information criterion BIC.In formula,
Wherein, as a line any in structure G is expressed as Lm→Ln, then KL
(Lm→Ln) it is relationship certainty value.Summation in formula is the document to the corresponding positive relationship of directed edge all in structure G
Knowledge confidence level is summed.
S32: non-directed graph structural constraint is constructed based on the relational knowledge base；For data-oriented collection D, to any in D
Illness is to LmAnd Ln, attribute is obtained to L to the relational knowledge base by retrieval illnessmAnd LnIllness to number, according to described
Illness is to illness in number searching document to LmAnd LnL in relationship certainty value tablem→LnRelationship certainty value and Ln→LmPass
It is certainty value.If L1Influence L2, then its connection relationship is L1Line L2And it is directed toward L2, construct L1With L2Directed edge, and assign just
To relationship certainty value.If L2Influence L1, then its connection relationship is L2Line L1And it is directed toward L2, construct L2With L1Directed edge, and assign
Give negative sense relationship certainty value.If L2Be independent of each other L1, then both not line, and assign vertical relation certainty value.
S33: Bayesian network is constructed based on the Bayesian network evaluation function and the non-directed graph structural constraint.It determines
After the non-directed graph structural constraint of Bayesian network, it is optimal to seek score function such as K2 algorithm for executable heuristic search algorithm
Network structure.General step are as follows: searched for since initial model, in each step of search, first with searching operators to current mould
Type carries out partial modification, obtains a series of candidate families, then calculates the scoring of each candidate family, and by best candidate model
Compared with "current" model.If the scoring of best candidate model is big, it is continued searching as next "current" model；Otherwise stop
It only searches for, returns to "current" model.According to bayesian principle, the maximum candidate family that scores is Bayesian network.Preferably, according to
Bayesian network evaluation function is constructed according to according to the Bayesian network and Bayes rule of foundation.Bayesian network evaluation function
It can be constructed according to classical heuristic structure learning algorithm, such as K2 algorithm, Max-Min Parents and Children are calculated
Method and Markov chain monte carlo search etc.
Cause and effect unit 3 is put down based on to excavate the average cause-effect between illness by data pattern so as to basis
Complication or complication whether are constituted between equal cause-effect illness.In average cause-effect, cause and effect unit 3 is based on Pearl
Principle and bayesian network structure calculate cause-effect average between illness.Whether Pearl is event exploring event X
When the reason of Y, needs to carry out X event by intervening X, calculate E (Y | do (X)), is i.e. event Y changes in the case where intervening X
Average case be greater than significance, then it is assumed that the reason of X is Y.Specifically, it in data-oriented collection D or Di, screens first
The illness for needing to study out, these illnesss include target conditions and the other illnesss for influencing the target conditions.For example, it is desired to grind
Study carefully illness L1 whether be illness L2 complication, the side of all illnesss for being directed toward L1 is truncated, observes illness L1 and illness L2 at this time
Average cause-effect, if it is this variation be greater than setting cause-effect threshold value, then it is assumed that illness L1 is constituted with illness L2
Complication, conversely, constituting complication.
When cause and effect unit 3 is based on to excavate the average cause-effect between illness by data pattern, due to Document Quantity
It is huge, to cause the huge of Bayes's grid, therefore, average cause-effect is calculated using back door criterion.Back door criterion is
Refer to, Bayes's grid G is a directed acyclic graph, (Lm, Ln) be G a pair of of node, node Z set be (Lm, Ln) back door,
Wherein, node all in Z is not the offspring of Z and Z has blocked all direction LmConnection LmTo LnPath.Therefore, may be used
With by back door principle come reasoning illness to LmAnd LnCausality.
Embodiment 2
Supplement of the present embodiment as embodiment 1.In order to the case where not influencing the causality between illness pair
Under, cause and effect unit 3 simplifies undirected constraint diagram by independence test.For example, independence test can be examined using card side's independence
It tests.
In the present invention, independence test can also be in the following way:
For illness Lm, the non-directed graph by way of compiling based on building obtains and LmThe node being connected constitutes its section
Point set.And gradually calculate each node and illness LmBetween correlation, and the node for therefrom choosing correlation maximum carry out it is only
Vertical property it is assumed that delete under given subset D with LmIndependent node.In the present invention, stochastic variable is measured to L using entropym's
It is uncertain.In given stochastic variable LmIn the case where, stochastic variable LnUncertainty can use conditional entropy for example under type measure:
Stochastic variable LnWith LmBetween degree of correlation can be measured by mutual information:
If mutual information has exceeded the threshold value of mutual information, then it is assumed that LnWith LmWith correlation.If mutual information without departing from
The threshold value of mutual information, then it is assumed that LnWith LmDo not have correlation.
Embodiment 3
The building system for the causality knowledge base towards medicine that present embodiment discloses a kind of, do not cause conflict or
In contradictory situation, the entirety and/or partial content of the preferred embodiment of other embodiments can be used as the benefit of the present embodiment
It fills.Preferably, this method can be realized by method of the invention and/or other alternative modules.
As shown in Fig. 2, the system mainly includes literary unit 1, data cell 2, cause and effect unit 3 and blocks of knowledge 4.Document
Unit 1 is configured in for constructing original library.Data cell is configured in 2 for constructing data set.Cause and effect unit 3 is matched
It is placed in for the average cause-effect between illness.Blocks of knowledge 4 is configured in for storing the original library, the number
According to collection and/or the average cause-effect to construct the knowledge base that can be read and/or show, in the form of data quantization
It is supplied to medical worker reference, study and/or decision.Preferably, literary unit 1 can be based on user-defined request
Numerous pertinent literatures containing various diseases simultaneously carry out classification to it and form several literary unit bodies to construct original library, with
Data cell 2 to be based on literary unit body to obtain main characteristic parameters and construct data set based on main characteristic parameters,
To reduce numerous characteristic parameters of numerous pertinent literatures formation for interference causal between illness and improve original
The utility value in library.Cause and effect unit 3 is based on main characteristic parameters and data set constructs Bayesian network, to be dug by data pattern
The average cause-effect between illness is dug, so as to whether constitute complication or merging according between average cause-effect illness
Disease.
Preferably, literary unit 1 counts the frequency of words/phrases in each document, obtains according to independence assumption single
Word/phrase joint probability of occurrence.Literary unit 1 calculates the relevance intensity of words/phrases, and is corrected based on relevance intensity
Joint probability of occurrence is to obtain the association reduced coordinate of the document.Association reduction of the literary unit 1 based on whole pertinent literatures
The classification function of coordinate building classifies pertinent literature according to the form of iterative algorithm to form several literary unit bodies.Its
In, classification function can carry out deep learning based on the sample size of pertinent literature, to enhance the precision of literary unit 1.
Preferably, in the case where data cell 2 obtains literary unit body, data cell 2 is according to illness to pairing
Mode obtains data set.Data cell 2 extracts wherein each pertinent literature in a manner of the syntactic analysis of natural language processing
Relationship between illness pair, to establish the relational knowledge base of illness pair, the relationship between illness pair includes positive relationship, reversed
Relationship and vertical relation.Also, data cell 2 is based on relationship knowledge table to the document for including illness pair in literary unit body
Retrieved relationship certainty value library of the relationship certainty value that illness pair is obtained in a manner of fusion to establish illness pair, illness pair
Between relationship include positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value.To 2 base of data cell
In to the relational knowledge base established in the way of matching two-by-two between all illnesss and relationship certainty value library building data set.
Cause and effect unit constructs Bayesian network as follows, S31: being commented based on relational knowledge base building Bayesian network
Valence function:
logP(G,D,KL)=logP (G)+log (D | G)+logP (KL|G)
S32: non-directed graph structural constraint is constructed based on relational knowledge base.To data set D, to any illness in data set D
To LmAnd Ln, attribute is obtained to L to relational knowledge base by retrieval illnessmAnd LnIllness to number, number is examined according to illness
Illness is to L in Suo WenxianmAnd LnL in relationship certainty value tablem→LnRelationship certainty value and Ln→LmRelationship certainty value.
S33: Bayesian network is constructed based on Bayesian network evaluation function and non-directed graph structural constraint.
Preferably, in the present invention, literary unit 1, data cell 2, cause and effect unit 3 and blocks of knowledge 4 are a kind of tools
There is the microprocessor of calculation function.For example, the literary unit 1 used in the present invention is that with search engine and have operation
The server of function.Data cell 2 is the data server with calculation function.Cause and effect unit 3 is the number with calculation function
According to server.Blocks of knowledge 4 is the memory with access facility, as RAM ROM disk at least one of cloud disk.Document
Unit 1, data cell 2, cause and effect unit 3 and blocks of knowledge 4 pass through optical fiber, data line, bluetooth, wifi and/or 4G between each other
It is connected etc. wired, wireless communication mode.
It should be noted that above-mentioned specific embodiment is exemplary, those skilled in the art can disclose in the present invention
Various solutions are found out under the inspiration of content, and these solutions also belong to disclosure of the invention range and fall into this hair
Within bright protection scope.It will be understood by those skilled in the art that description of the invention and its attached drawing are illustrative and are not
Constitute limitations on claims.Protection scope of the present invention is defined by the claims and their equivalents.
Claims (10)
1. a kind of building system of the causality knowledge base towards medicine, comprising:
Literary unit (1): for constructing original library；
Data cell (2): for constructing data set；
Cause and effect unit (3): for calculating the average cause-effect between illness；
Blocks of knowledge (4): for storing the original library, the data set and/or the average cause-effect to construct
The knowledge base that can be read and/or show,
It is characterized in that,
The literary unit (1), which can obtain numerous pertinent literatures containing various diseases and carry out classification to it, forms several texts
Cell cube is offered to construct original library, so that the data cell (2) can be obtained mainly based on the literary unit body
Characteristic parameter simultaneously constructs data set based on the main characteristic parameters,
The cause and effect unit (3) is based on the main characteristic parameters and the data set constructs Bayesian network, to pass through data
Average cause-effect between pattern analysis illness, so that the blocks of knowledge (4) can be based on the pertinent literature to be formed
The mode of the corresponding relationship of the average cause-effect between the illness constructs the knowledge base.
2. building system as described in claim 1, which is characterized in that the literary unit (1) is configured in for counting every
The frequency of words/phrases in one document obtains the word/phrase joint probability of occurrence according to independence assumption；
The literary unit (1) calculates the relevance intensity of the word/phrase, and is corrected based on the relevance intensity
It is described to combine probability of occurrence to obtain the association reduced coordinate of the document；
The literary unit (1) literary unit (1) constructs the association reduced coordinate of the document, and based on described in whole
The classification function that the association reduced coordinate of pertinent literature and the relevance intensity construct is by the pertinent literature according to repeatedly
Classify for the form of algorithm to form several literary unit bodies；
Wherein, the classification function can carry out deep learning based on the sample size of the pertinent literature, to enhance the text
Offer the precision of unit (1).
3. building system as claimed in claim 1 or 2, which is characterized in that obtain the document in the data cell (2)
In the case where cell cube, the data cell (2) is configured in obtains the data set in such a way that illness is to pairing:
The data cell (2) extracts each described pertinent literature wherein sick in a manner of the syntactic analysis of natural language processing
Relationship between disease pair, to establish the relational knowledge base of the illness pair, the relationship between the illness pair includes positive closes
System, inverse relationship and vertical relation；
Also, the data cell (2) is based on relationship knowledge table to including the illness pair in the literary unit body
Document is retrieved the relationship certainty value for being obtained the illness pair in a manner of fusion and is believed to establish the relationship of the illness pair
Angle value library, the relationship between the illness pair includes positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value；
To which, the data cell (2) is based on knowing the relationship established in the way of matching two-by-two between all illnesss
Know library and relationship certainty value library constructs the data set.
4. the building system as described in one of preceding claims, which is characterized in that the cause and effect unit (3) be configured according to
If under type constructs Bayesian network,
S31: Bayesian network evaluation function is constructed based on the relational knowledge base:
logP(G,D,KL)=logP (G)+log (D | G)+logP (KL|G)
S32: non-directed graph structural constraint is constructed based on the relational knowledge base；To the data set D, in the data set D
Any illness is to LmAnd Ln, attribute is obtained to L to the relational knowledge base by retrieval illnessmAnd LnIllness to number, according to
The illness is to illness in number searching document to LmAnd LnL in relationship certainty value tablem→LnRelationship certainty value and Ln→Lm
Relationship certainty value,
S33: Bayesian network is constructed based on the Bayesian network evaluation function and the non-directed graph structural constraint.
5. the building system as described in one of preceding claims, which is characterized in that the cause and effect unit (3), which is configured in, to be based on
The Bayesian network and Pearl principle calculate the average cause-effect between each illness pair,
When the average cause-effect is more than the cause-effect threshold value of setting, complication is constituted between illness；Described average
When cause-effect is no more than the cause-effect threshold value of setting, complication is constituted between illness.
6. the building system as described in one of preceding claims, which is characterized in that for illness Lm, the base by way of traversal
In the non-directed graph structural constraint acquisition and illness LmThe node being connected constitutes its node collection；And gradually calculate each node with
Illness LmBetween correlation, and therefrom choose correlation maximum node carry out independence assumption, delete in data-oriented collection D
Lower and LmIndependent node；
Illness LnWith illness LmBetween independence pass through mutual information measure:
When the mutual information has exceeded the threshold value of mutual information, then illness LnWith illness LmIt is less independent with correlation；Institute
The threshold value of the mutual information without departing from mutual information is stated, then illness LnWith illness LmIt is independent without correlation.
7. a kind of construction method of the causality knowledge base towards medicine, is related to a kind of a variety of diseases of the excavation based on literature search
Whether the construction method of the knowledge base of complication or complication is constituted between disease, and step includes:
Literary unit (1) constructs original library；
Data cell (2) constructs data set；
Cause and effect unit (3) is used to calculate the average cause-effect between illness；
Blocks of knowledge (4) stores the original library, the data set and/or the average cause-effect and can be read with constructing
The knowledge base for taking and/or showing；
It is characterized in that,
The literary unit (1), which can obtain numerous pertinent literatures containing various diseases and carry out classification to it, forms several texts
Cell cube is offered to construct original library, so that the data cell (2) can be obtained mainly based on the literary unit body
Characteristic parameter simultaneously constructs data set based on the main characteristic parameters,
The cause and effect unit (3) is based on the main characteristic parameters and the data set constructs Bayesian network, to pass through data
Average cause-effect between pattern analysis illness, so that the blocks of knowledge (4) can be based on the pertinent literature to be formed
The mode of the corresponding relationship of the average cause-effect between the illness and between the illness constructs the knowledge base.
8. construction method as claimed in claim 7, which is characterized in that the pertinent literature classification is carried out as follows:
S11: the literary unit (1) counts the frequency of words/phrases in each document, obtain according to independence assumption described in
Word/phrase joint probability of occurrence；
S12: the literary unit (1) calculates the relevance intensity of the word/phrase；
S13: the literary unit (1) constructs the association reduced coordinate of the document, and based on whole pertinent literatures
The classification function that the association reduced coordinate and the relevance intensity construct is by the pertinent literature according to the shape of iterative algorithm
Formula is classified to form several literary unit bodies.
9. construction method as claimed in claim 7 or 8, which is characterized in that obtain the document in the data cell (2)
In the case where cell cube, the data cell (2) obtains the data in such a way that illness forms the illness pair to pairing
Collection；
The data cell (2) extracts the disease each described pertinent literature in a manner of the syntactic analysis of natural language processing
Relationship between disease pair, to establish the relational knowledge base of the illness pair, the relationship between the illness pair includes positive closes
System, inverse relationship and vertical relation；
Also, the data cell (2) is based on relationship knowledge table to including the illness pair in the literary unit body
Document is retrieved the relationship certainty value for being obtained the illness pair in a manner of fusion and is believed to establish the relationship of the illness pair
Angle value library, the relationship between the illness pair includes positive relationship certainty value, inverse relationship certainty value and vertical relation certainty value；
To which, the data cell (2) is based on knowing the relationship established in the way of matching two-by-two between all illnesss
Know library and relationship certainty value library constructs the data set D.
10. the construction method as described in one of claim 7 to 9, which is characterized in that the cause and effect unit (3) is according to such as lower section
Formula constructs Bayesian network,
S31: Bayesian network evaluation function is constructed based on the relational knowledge base:
logP(G,D,KL)=logP (G)+logP (D | G)+logP (KL|G)
S32: non-directed graph structural constraint is constructed based on the relational knowledge base；For the data set D, in the data set D
Any illness to LmAnd Ln, attribute is obtained to L to the relational knowledge base by retrieval illnessmAnd LnIllness to number, root
According to the illness to illness in number searching document to LmAnd LnL in relationship certainty value tablem→LnRelationship certainty value and Ln→
LmRelationship certainty value,
S33: Bayesian network is constructed based on the Bayesian network evaluation function and the non-directed graph structural constraint.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011106913.3A CN112151130B (en) | 2019-01-15 | 2019-01-15 | Decision support system based on literature retrieval and construction method |
| CN201910034537.2A CN109801687B (en) | 2019-01-15 | 2019-01-15 | Health field-oriented cause and effect relationship knowledge base construction method and system |
| CN202011106914.8A CN112233736B (en) | 2019-01-15 | 2019-01-15 | Knowledge base construction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910034537.2A CN109801687B (en) | 2019-01-15 | 2019-01-15 | Health field-oriented cause and effect relationship knowledge base construction method and system |
Related Child Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011106914.8A Division CN112233736B (en) | 2019-01-15 | 2019-01-15 | Knowledge base construction method and system |
| CN202011106913.3A Division CN112151130B (en) | 2019-01-15 | 2019-01-15 | Decision support system based on literature retrieval and construction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109801687A true CN109801687A (en) | 2019-05-24 |
| CN109801687B CN109801687B (en) | 2020-11-10 |
Family
ID=66558921
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011106914.8A Active CN112233736B (en) | 2019-01-15 | 2019-01-15 | Knowledge base construction method and system |
| CN202011106913.3A Active CN112151130B (en) | 2019-01-15 | 2019-01-15 | Decision support system based on literature retrieval and construction method |
| CN201910034537.2A Active CN109801687B (en) | 2019-01-15 | 2019-01-15 | Health field-oriented cause and effect relationship knowledge base construction method and system |
Family Applications Before (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011106914.8A Active CN112233736B (en) | 2019-01-15 | 2019-01-15 | Knowledge base construction method and system |
| CN202011106913.3A Active CN112151130B (en) | 2019-01-15 | 2019-01-15 | Decision support system based on literature retrieval and construction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN112233736B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111949792A (en) * | 2020-08-13 | 2020-11-17 | 电子科技大学 | Medicine relation extraction method based on deep learning |
| CN112463973A (en) * | 2019-09-06 | 2021-03-09 | 医渡云（北京）技术有限公司 | Construction method, device and medium of medical knowledge graph and electronic equipment |
| CN114664452A (en) * | 2022-05-20 | 2022-06-24 | 之江实验室 | General multi-disease prediction system based on causal verification data generation |
| CN114864099A (en) * | 2022-07-05 | 2022-08-05 | 浙江大学 | Clinical data automatic generation method and system based on causal relationship mining |
| CN117891959A (en) * | 2024-03-15 | 2024-04-16 | 中国标准化研究院 | Document metadata storage method and system based on Bayesian network |
| CN117891959B (en) * | 2024-03-15 | 2024-05-10 | 中国标准化研究院 | Document metadata storage method and system based on Bayesian network |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113538178A (en) * | 2021-06-10 | 2021-10-22 | 北京易创新科信息技术有限公司 | Intellectual property value evaluation method and device, electronic equipment and readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763528A (en) * | 2009-12-25 | 2010-06-30 | 深圳大学 | Gene regulation and control network constructing method based on Bayesian network |
| CN103493054A (en) * | 2010-10-12 | 2014-01-01 | 美国西门子医疗解决公司 | Healthcare information technology system for predicting development of cardiovascular conditions |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7644053B2 (en) * | 2004-03-03 | 2010-01-05 | The Boeing Company | System, method, and computer program product for combination of cognitive causal models with reasoning and text processing for knowledge driven decision support |
| CN102855398B (en) * | 2012-08-28 | 2016-03-02 | 中国科学院自动化研究所 | The acquisition methods of the potential associated gene of the disease based on Multi-source Information Fusion |
| WO2016040725A1 (en) * | 2014-09-11 | 2016-03-17 | Berg Llc | Bayesian causal relationship network models for healthcare diagnosis and treatment based on patient data |
| CN104361033B (en) * | 2014-10-27 | 2017-06-09 | 深圳职业技术学院 | A kind of automatic collection method of cancer relevant information and system |
| CN106295187A (en) * | 2016-08-11 | 2017-01-04 | 中国科学院计算技术研究所 | Construction of knowledge base method and system towards intelligent clinical auxiliary decision-making support system |
| CN106667443A (en) * | 2017-01-10 | 2017-05-17 | 中山大学中山眼科中心 | Congenital cataract postoperative complication predicting method and system |
| CN107145712B (en) * | 2017-04-06 | 2021-01-01 | 广州慧扬健康科技有限公司 | Medical record statistical analysis system for complications and complicating diseases |
| CN107887036A (en) * | 2017-11-09 | 2018-04-06 | 北京纽伦智能科技有限公司 | Construction method, device and the clinical decision accessory system of clinical decision accessory system |
| CN108986871A (en) * | 2018-08-27 | 2018-12-11 | 东北大学 | A kind of construction method of intelligent medical treatment knowledge mapping |
-
2019
- 2019-01-15 CN CN202011106914.8A patent/CN112233736B/en active Active
- 2019-01-15 CN CN202011106913.3A patent/CN112151130B/en active Active
- 2019-01-15 CN CN201910034537.2A patent/CN109801687B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763528A (en) * | 2009-12-25 | 2010-06-30 | 深圳大学 | Gene regulation and control network constructing method based on Bayesian network |
| CN103493054A (en) * | 2010-10-12 | 2014-01-01 | 美国西门子医疗解决公司 | Healthcare information technology system for predicting development of cardiovascular conditions |
Non-Patent Citations (5)
| Title |
|---|
| YIHENG LIANG等: ""Big Data Problems on Discovering and Analyzing Causal Relationships in Epidemiological Data"", 《2014 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA)》 * |
| 安宁等: ""基于因果效应的贝叶斯网络结构学习方法"", 《计算机应用研究》 * |
| 张润梅: ""基于贝叶斯网络的复杂系统因果关系研究"", 《中国博士学位论文全文数据库 基础科学辑》 * |
| 王至婉等: ""基于贝叶斯网络构建慢性阻塞性肺疾病急性加重期痰热壅肺证症状与证候间关联模式"", 《中医杂志》 * |
| 范建平等: ""贝叶斯网络在中医诊断中的应用研究"", 《管理科学学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112463973A (en) * | 2019-09-06 | 2021-03-09 | 医渡云（北京）技术有限公司 | Construction method, device and medium of medical knowledge graph and electronic equipment |
| CN111949792A (en) * | 2020-08-13 | 2020-11-17 | 电子科技大学 | Medicine relation extraction method based on deep learning |
| CN111949792B (en) * | 2020-08-13 | 2022-05-31 | 电子科技大学 | Medicine relation extraction method based on deep learning |
| CN114664452A (en) * | 2022-05-20 | 2022-06-24 | 之江实验室 | General multi-disease prediction system based on causal verification data generation |
| CN114664452B (en) * | 2022-05-20 | 2022-09-23 | 之江实验室 | General multi-disease prediction system based on causal verification data generation |
| CN114864099A (en) * | 2022-07-05 | 2022-08-05 | 浙江大学 | Clinical data automatic generation method and system based on causal relationship mining |
| CN117891959A (en) * | 2024-03-15 | 2024-04-16 | 中国标准化研究院 | Document metadata storage method and system based on Bayesian network |
| CN117891959B (en) * | 2024-03-15 | 2024-05-10 | 中国标准化研究院 | Document metadata storage method and system based on Bayesian network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112151130A (en) | 2020-12-29 |
| CN109801687B (en) | 2020-11-10 |
| CN112233736B (en) | 2022-09-16 |
| CN112151130B (en) | 2022-11-04 |
| CN112233736A (en) | 2021-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109801687A (en) | A kind of construction method and system of the causality knowledge base towards medicine | |
| Abbas et al. | A literature review on the state-of-the-art in patent analysis | |
| Antal et al. | Using literature and data to learn Bayesian networks as clinical models of ovarian tumors | |
| Fang et al. | Feature Selection Method Based on Class Discriminative Degree for Intelligent Medical Diagnosis. | |
| Alpan et al. | Classification of diabetes dataset with data mining techniques by using WEKA approach | |
| CN105260598A (en) | Oral diagnosis and treatment decision support system and decision method | |
| Shen et al. | Discovering the potential opportunities of scientific advancement and technological innovation: A case study of smart health monitoring technology | |
| Badlani et al. | Multilingual healthcare chatbot using machine learning | |
| Abdikenov et al. | Analytics of heterogeneous breast cancer data using neuroevolution | |
| Lee et al. | Automatic generation of concept hierarchies using WordNet | |
| Chen et al. | Unite: Uncertainty-based health risk prediction leveraging multi-sourced data | |
| Wenxin | Heart disease prediction model based on model ensemble | |
| CN109800881A (en) | A kind of causality determination method based on Bayesian network | |
| Sharma et al. | A diabetes monitoring system and health-medical service composition model in cloud environment | |
| Jeon et al. | Measuring the novelty of scientific publications: A fastText and local outlier factor approach | |
| Wegrzyn-Wolska et al. | Social media analysis for e-health and medical purposes | |
| Rijcken et al. | Exploring embedding spaces for more coherent topic modeling in electronic health records | |
| Das et al. | Graph-based text summarization and its application on COVID-19 twitter data | |
| Wang et al. | TransH-RA: A Learning Model of Knowledge Representation by Hyperplane Projection and Relational Attributes | |
| Petrović | Tackling the COVID-19 conspiracies: the data-driven approach | |
| Wang et al. | Non-greedy tree-based learning for estimating global optimal dynamic treatment decision rules with continuous treatment dosage | |
| Singh et al. | Comparative analysis of data mining models for classification for small data set | |
| Shah et al. | Comparison of deep learning based concept representations for biomedical document clustering | |
| Phuc et al. | Using SOM based graph clustering for extracting main ideas from documents | |
| Surve et al. | HealthCare Chatbot Using Machine Learning and NLP |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230511 Address after: Room 215-E, 2nd Floor, Podium Building, No. 33 Yueming Road, Xixing Street, Binjiang District, Hangzhou City, Zhejiang Province, 310000 China (Zhejiang) Free Trade Pilot Zone Patentee after: Hangzhou Zhilan health Co.,Ltd. Address before: 230009 No. 193, Tunxi Road, Hefei, Anhui Patentee before: Hefei University of Technology |
|
| TR01 | Transfer of patent right |