CN110929040A - Knowledge graph construction method and device for specific medical field - Google Patents
Knowledge graph construction method and device for specific medical field Download PDFInfo
- Publication number
- CN110929040A CN110929040A CN201911046367.6A CN201911046367A CN110929040A CN 110929040 A CN110929040 A CN 110929040A CN 201911046367 A CN201911046367 A CN 201911046367A CN 110929040 A CN110929040 A CN 110929040A
- Authority
- CN
- China
- Prior art keywords
- medical
- knowledge graph
- knowledge
- constructing
- medical field
- 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
Images
Classifications
-
- 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
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
-
- 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
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- 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
Abstract
The invention provides a knowledge graph construction method and a device aiming at a specific medical field, wherein the method comprises the following steps: constructing a concept knowledge graph of the target medical field according to medical standards in a medical word bank and prior knowledge of doctors in the target medical field; constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field; combining the concept knowledge graph and the instance knowledge graph into a fact knowledge graph. The invention uses the prior knowledge of doctors to identify the concepts and relations in the specific medical field, so that the constructed knowledge graph is more in line with the actual requirements; meanwhile, the embodiment can construct knowledge maps of various different medical fields, and can be expanded to other disease fields according to prior knowledge and actual requirements of doctors based on the constructed knowledge maps of the specific medical fields.
Description
Technical Field
The invention belongs to the technical field of medical knowledge maps, and particularly relates to a knowledge map construction method and device for the specific medical field.
Background
With the development of medical informatization and the large-scale increase of the clinical information data volume, the knowledge graph plays an increasingly important role in the medical field. The medical field is highly specialized, has rich and diverse high-quality medical concept resources, and has many specialized sub-fields, such as the cardiovascular disease field, the diabetes field, and the like. Due to the difficulty of data acquisition and the lack of a priori knowledge from the disease domain, it is very difficult to construct a medical knowledge map covering all disease domains at once.
The construction system architecture of the conventional knowledge graph mainly comprises the following contents: knowledge representation, knowledge graph construction tools, such as information extraction and fusion tools, and knowledge storage and application. However, conventional strategies cannot be directly applied to the construction of domain-specific profiles, let alone more specialized disease-domain-specific-oriented profiles. Furthermore, in the medical field, there are various diseases, many complex concepts and relationships, which are difficult to recognize using only a medical thesaurus.
In summary, the conventional knowledge graph construction method cannot be directly applied to the construction of the knowledge graph in the specific medical field. Therefore, it is highly desirable to provide a method for constructing a knowledge map for a specific disease domain.
Disclosure of Invention
In order to overcome the problem that the conventional knowledge graph construction method cannot be directly applied to the construction of the knowledge graph in the specific medical field or at least partially solve the problem, the embodiment of the invention provides a knowledge graph construction method and a knowledge graph construction device for the specific medical field.
According to a first aspect of the embodiments of the present invention, there is provided a method for constructing a knowledge graph for a specific disease domain, including:
constructing a concept knowledge graph of the target medical field according to medical standards in a medical word bank and prior knowledge of doctors in the target medical field;
constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field;
combining the concept knowledge graph and the instance knowledge graph into a fact knowledge graph.
Preferably, the step of constructing a conceptual knowledge-graph of the target medical field based on the medical criteria in the medical thesaurus and the prior knowledge of the doctor in the target medical field comprises:
constructing a medical synonym library and a general concept knowledge map according to medical standards in the medical thesaurus;
and constructing the concept knowledge graph of the target medical field according to the medical synonym library, the general concept knowledge graph and the prior knowledge of the doctor of the target medical field.
Preferably, the step of constructing the medical thesaurus and the generic concept knowledge graph according to medical criteria in the medical thesaurus comprises:
mapping the medical word stock from an ER data model to an RDF data model according to a mapping method which is selected by the doctor for the medical word stock in advance;
and constructing the medical synonym library and the universal concept knowledge map according to the medical thesaurus of the RDF data model.
Preferably, the step of constructing a conceptual knowledge-graph of the target medical field according to the medical criteria in the medical thesaurus and the prior knowledge of the doctor in the target medical field further comprises:
acquiring the concept and the relation input by the doctor, and representing the concept and the relation by a triple in an RDF data model;
judging whether the input of the doctor exists in the medical word stock and the concept knowledge graph of the target medical field;
if not, inserting the input into the medical thesaurus and the concept knowledge graph of the target medical field;
if so, acquiring standard medical knowledge corresponding to the input according to the medical standard, and inserting the standard medical knowledge into the concept knowledge graph of the target medical field.
Preferably, the step of constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field comprises:
according to a mapping method which is selected by the doctor for the electronic medical record in advance, mapping the structured data in the electronic medical record from an ER data model to an RDF data model;
extracting entities, events and relations from unstructured data of the electronic medical record by adopting a machine learning or deep learning method according to the labels of doctors in the electronic medical record;
the structured data refers to data which has a certain structure and is managed in a relational database table form; unstructured data refers to data without fixed structural patterns, such as paragraphs in a corpus, etc.;
and generating the instance knowledge graph according to the structured data of the EDF data model and the extracted entities, events and relations.
Preferably, the mapping method is D2R, R2RML or Virtuoso.
Preferably, the medical thesaurus is a UMLS medical thesaurus.
According to a second aspect of the embodiments of the present invention, there is provided a knowledge graph constructing apparatus for a specific disease domain, including:
the first construction module is used for constructing a concept knowledge graph of the target medical field according to the medical standard and the prior knowledge of doctors in the target medical field;
the second construction module is used for constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field;
a combining module for combining the concept knowledge graph and the instance knowledge graph into a fact knowledge graph.
According to a third aspect of the embodiments of the present invention, there is also provided an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor calls the program instructions to execute the method for constructing a knowledge graph for a specific disease domain provided in any one of the various possible implementations of the first aspect.
According to a fourth aspect of embodiments of the present invention, there is also provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform a method of constructing a knowledge graph for a specific disease domain as provided by any one of the various possible implementations of the first aspect.
The embodiment of the invention provides a method and a device for constructing a knowledge graph aiming at a specific medical field, wherein the method is used for constructing the medical knowledge graph aiming at the specific medical field by using the prior knowledge of a doctor to identify the concept and the relation of the specific medical field based on the existing medical word stock, the electronic medical record of a professional hospital from the specific medical field and an expert for researching the disease, so that the constructed knowledge graph is more in line with the actual requirement; meanwhile, the embodiment can construct knowledge maps of various different medical fields, and can be expanded to other disease fields according to prior knowledge and actual requirements of doctors based on the constructed knowledge maps of the specific medical fields.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a flow chart of a method for constructing a knowledge graph for a specific medical field according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a conceptual knowledge graph building process in a knowledge graph building method for a specific medical field according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating an example knowledge graph construction process in a knowledge graph construction method for a specific medical field according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of the overall construction process of the concept knowledge-graph and the example knowledge-graph in the knowledge-graph construction method for a specific medical field according to the embodiment of the present invention;
FIG. 5 is a schematic diagram of the overall architecture of the concept knowledge-graph and the example knowledge-graph construction in the knowledge-graph construction method for the specific medical field according to the embodiment of the present invention;
FIG. 6 is a schematic diagram of an example knowledge graph structure of a diabetes domain in a knowledge graph construction method for a specific medical domain according to an embodiment of the present invention;
FIG. 7 is a schematic structural diagram of a knowledge graph constructing apparatus for a specific medical field according to an embodiment of the present invention;
fig. 8 is a schematic view of an overall structure of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
In an embodiment of the present invention, a method for constructing a knowledge graph for a specific medical field is provided, and fig. 1 is an overall flowchart of the method for constructing a knowledge graph for a specific medical field according to an embodiment of the present invention, where the method includes: s101, constructing a concept knowledge graph of the target medical field according to medical standards in a medical word bank and prior knowledge of doctors of the target medical field;
the Medical word stock may be a Medical word stock containing different Medical standards, such as UMLS (Unified Medical Language System). The concept knowledge graph in this embodiment includes concepts and relationships between the concepts. The medical thesaurus contains a large number of medically related concepts, for example, clinical, basic, pharmaceutical, biological, medical management and other medical concepts are covered in UMLS. Since the embodiment requires the construction of a knowledge graph of a specific medical field, such as the medical field of heart disease, the concepts and relationships related to the specific medical field are selected from the medical lexicon depending on the prior knowledge of the doctor, and the concept knowledge graph of the specific medical field is constructed according to the medical standards in the medical lexicon.
S102, constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field;
among them, the electronic medical records of the target medical field are clinical data from a professional hospital for treating a specific disease. The instance knowledge graph comprises entities, events and relations, and the relations comprise relations between the entities and the events. And extracting the information required by the example knowledge graph from the electronic medical record in the target medical field to construct the example knowledge graph.
S103, combining the concept knowledge graph and the example knowledge graph into a fact knowledge graph.
And fusing concept nodes in the concept knowledge graph and example nodes in the example knowledge graph to generate a fact knowledge graph, wherein the fact knowledge graph comprises concepts, entities, events and relations.
In the embodiment, the medical knowledge graph in the specific disease field is constructed based on the existing medical word stock, the electronic medical record of the professional hospital from the specific disease field and the expert for researching the disease, and the concept and the relation of the specific medical field are identified by using the prior knowledge of a doctor, so that the constructed knowledge graph is more in line with the actual requirement; meanwhile, the embodiment can construct knowledge maps of various different medical fields, and can be expanded to other disease fields according to prior knowledge and actual requirements of doctors based on the constructed knowledge maps of the specific medical fields.
On the basis of the above embodiment, in this embodiment, the step of constructing the conceptual knowledge graph of the target medical field according to the medical standard in the medical thesaurus and the prior knowledge of the doctor in the target medical field includes: constructing a medical synonym library and a general concept knowledge map according to medical standards in the medical thesaurus; and constructing the concept knowledge graph of the target medical field according to the medical synonym library, the general concept knowledge graph and the prior knowledge of the doctor of the target medical field.
Specifically, in order to more quickly and effectively construct the concept knowledge graph of the target medical field, the medical synonym library and the general concept knowledge graph are firstly constructed according to the medical standard, namely the medical synonym library and the general concept knowledge graph are firstly constructed according to all contents in the medical thesaurus by using the medical standard. Wherein, the general concept knowledge graph refers to a concept knowledge graph covering a plurality of fields. And then constructing a concept knowledge graph facing the specific disease field according to the prior knowledge of doctors in the specific medical field and the synonym library and the general concept knowledge graph. Specifically, prior knowledge input by a doctor is obtained, the prior knowledge comprises concepts and relations, synonyms corresponding to the doctor input are obtained from a synonym library, information related to the synonyms is obtained from a general concept knowledge map according to the synonyms corresponding to the doctor input, and a concept knowledge map in a specific disease field is constructed according to the doctor input and the information related to the synonyms.
On the basis of the above embodiment, the step of constructing the medical synonym library and the generic concept knowledge base according to the medical standard in the medical thesaurus in this embodiment includes: mapping the medical word stock from an ER (Entity Relationship) data model to an RDF (resource description Framework) data model according to a mapping method which is selected by the doctor for the medical word stock in advance; and constructing the medical synonym library and the universal concept knowledge map according to the medical thesaurus of the RDF data model.
Specifically, the present embodiment provides a rule base tool for a physician to select a mapping method from an ER structure model stored in a medical thesaurus, such as UMLS, to an RDF model to be used, the mapping method being used to convert data from the ER data model to the RDF data model, as shown in fig. 2. The mapping is then performed using the mapping tools and mapping methods stored in the rule base tool, according to the physician's selection. Mapping tools and mapping methods such as D2R, R2RML, Virtuoso, etc., which are not limited by the present embodiment. And then the prior knowledge input by the doctor through the input component is normalized through a generalization tool and then added into the concept knowledge graph.
On the basis of the foregoing embodiment, in this embodiment, the step of constructing the concept knowledge graph of the target medical field according to the medical synonym library, the general concept knowledge graph and the prior knowledge of the doctor in the target medical field further includes: acquiring the concept and the relation input by the doctor, and representing the concept and the relation by a triple in an RDF data model; judging whether the input of the doctor exists in the medical word stock and the concept knowledge graph of the target medical field; if not, inserting the input into the medical thesaurus and the concept knowledge graph of the target medical field; if so, acquiring standard medical knowledge corresponding to the input according to the medical standard, and inserting the standard medical knowledge into the concept knowledge graph of the target medical field.
Specifically, the conceptual knowledge base of a specific disease in the present embodiment is constructed according to the prior knowledge of the doctor and the medical lexicon. In order to obtain the prior knowledge of the doctor, the doctor defines the probabilities and relationships of the new disease domain by means of an input tool and adds them to the conceptual knowledge-graph of the target domain. The method comprises the following specific steps: providing relevant medical knowledge by a doctor in the field of new diseases, wherein the medical knowledge consists of medical concepts and relations and is represented by using triples; the knowledge of the doctor is standardized and persisted. The process of normalization and persistence falls into two categories depending on whether new knowledge exists in the medical thesaurus and the conceptual knowledge map of the target medical domain. If the UMLS does not exist, inserting the UMLS into a medical word stock and a concept knowledge graph of the target medical field according to the coding system of the UMLS; if so, standard medical knowledge encoded using UMLS is inserted into the conceptual knowledge-graph of the target medical domain.
And constructing a comprehensive medical word stock across disease categories step by step based on the steps. The core of the method is to construct a special medical word stock facing to a specific disease field based on a general medical word stock so as to meet related requirements.
On the basis of the above embodiment, in this embodiment, the step of constructing the example knowledge graph of the target medical field according to the electronic medical record of the target medical field includes: according to a mapping method which is selected by the doctor for the electronic medical record in advance, mapping the structured data in the electronic medical record from an ER data model to an RDF data model; extracting entities, events and relations from unstructured data of the electronic medical record by adopting a machine learning or deep learning method according to the labels of doctors in the electronic medical record; and generating the instance knowledge graph according to the structured data of the EDF data model and the extracted entities, events and relations.
Specifically, as shown in fig. 3, in order to better improve the data processing capability of the large-batch electronic medical records, the embodiment provides a rule base tool and a doctor labeling tool. Meanwhile, the embodiment also provides a plurality of selectable extraction methods based on machine learning and deep learning for the application scene for the user. The physician annotation tool provides a relevant interface through which the physician can annotate unstructured data according to the conceptual knowledge map. The annotated result is stored in the entity corpus and the relation corpus, so that the establishment of the entity and relation extraction model is supported.
As shown in FIG. 4, the data in the electronic medical record includes structured data and unstructured data. The method comprises the steps that structured data in the electronic medical record are generally an ER data model, the structured data are mapped to an RDF data model from the ER data model according to a mapping method selected by a doctor in advance, and an example knowledge graph is constructed; meanwhile, a doctor labels the electronic medical record through a labeling tool, extracts entities, events and relations from unstructured data of the electronic medical record by adopting a machine learning or deep learning method, generates another example knowledge graph, and fuses the two example knowledge graphs into an example knowledge graph in a specific medical field.
As shown in fig. 5, the data sources in this embodiment are doctor requirements, doctor prior knowledge, electronic medical records from hospitals, and an integrated medical language system, and in order to participate in the expert, it is necessary to provide tools for assisting the doctor in constructing the knowledge graph, i.e. a doctor input tool, a doctor labeling tool, and a rule base tool. And constructing a concept knowledge graph on the doctor input tool and the rule base tool, and constructing an example knowledge graph on the basis of the doctor marking tool and the rule base tool. The construction of the concept knowledge map is based on the prior knowledge of doctors and a medical word stock, and the module comprises a generalization tool, an ER-OWL mapping tool and a doctor input tool. The construction of the example knowledge graph is based on the electronic medical record, the labeling tool and the ER-OWL mapping tool of the hospital. Finally, the present embodiment will generate both a concept and instance knowledge-graph.
In addition to building a domain-specific disease-oriented knowledge map, the present embodiments can also incorporate new disease domains into the current knowledge map. The constructed medical synonym library and the concept knowledge graph do not need to be constructed again, except for the input of the electronic medical record and the prior knowledge of the doctor, the doctor participation platform and other construction tools can be repeatedly used, so that the workload of constructing the knowledge graph aiming at another specific disease field is reduced, and the knowledge graph covering different types of disease fields can be constructed step by step.
The following procedure is performed, taking the example of incorporating diabetes into an existing cardiovascular disease knowledge map:
first, there is a need to acquire electronic medical record data and doctors in the diabetes field. Physicians need to provide their prior knowledge from three aspects: 1) diabetes-related concepts, relationships, and RDF triplets; 2) mapping the structured data in the electronic medical record from the ER data model to the related rules of the RDF data model; 3) and (5) making an entity and relation extraction rule in unstructured data in the electronic medical record.
Then, a conceptual knowledge map of diabetes would be constructed. The prior knowledge of a doctor is obtained through a relevant input tool, and then the concept knowledge graph in the diabetes field is constructed based on the existing medical word stock, the relevant concept knowledge graph and the relevant concepts, relations and triples in the diabetes field. Establishing an example knowledge map of the diabetes domain is accomplished as shown in fig. 6. In the process, firstly, a doctor labels unstructured data in an electronic medical record based on a conceptual knowledge graph and related requirements in the field of diabetes; secondly, generating a relevant extraction model by adopting a machine learning or deep learning method based on the labeling result; extracting entities and relations of unstructured data in the electronic medical records in a large batch based on the extraction model, and generating a related corpus at the same time; and finally, aligning the extracted entities and the relations based on the concept knowledge graph and finally completing the construction of the instance knowledge graph.
In another embodiment of the present invention, a knowledge graph constructing apparatus for a specific medical field is provided, which is used for implementing the method in the foregoing embodiments. Therefore, the descriptions and definitions in the embodiments of the aforementioned knowledge graph construction method for a specific medical field can be used for understanding the various execution modules in the embodiments of the present invention. Fig. 7 is a schematic overall structure diagram of a knowledge graph constructing apparatus for a specific medical field according to an embodiment of the present invention, the apparatus includes a first constructing module 701, a second constructing module 702, and a combining module 703, wherein:
the first construction module 701 is used for constructing a concept knowledge graph of the target medical field according to the medical standard and the prior knowledge of doctors in the target medical field;
the medical word stock can be UMLS medical word stock containing different medical standards. The concept knowledge graph in this embodiment includes concepts and relationships between the concepts. The medical thesaurus contains a large number of medically relevant concepts. Since the embodiment requires to construct a knowledge graph of a specific medical field, the first construction module 701 selects concepts and relationships related to the specific medical field from the medical lexicon depending on prior knowledge of the doctor, and constructs the concept knowledge graph of the specific medical field according to medical standards in the medical lexicon.
The second construction module 702 is configured to construct an example knowledge graph of a target medical field according to an electronic medical record of the target medical field;
among them, the electronic medical records of the target medical field are clinical data from a professional hospital for treating a specific disease. The instance knowledge graph comprises entities, events and relations, and the relations comprise relations between the entities and the events. The second construction module 702 extracts the information required by the example knowledge graph from the electronic medical records in the target medical field to construct the example knowledge graph.
The combining module 703 is configured to combine the concept knowledge-graph and the instance knowledge-graph into a fact knowledge-graph.
The combination module 703 fuses concept nodes in the concept knowledge graph and instance nodes in the instance knowledge graph to generate a fact knowledge graph, where the fact knowledge graph includes concepts, entities, events and relationships.
In the embodiment, the medical knowledge graph in the specific disease field is constructed based on the existing medical word stock, the electronic medical record of the professional hospital from the specific disease field and the expert for researching the disease, and the concept and the relation of the specific medical field are identified by using the prior knowledge of a doctor, so that the constructed knowledge graph is more in line with the actual requirement; meanwhile, the embodiment can construct knowledge maps of various different medical fields, and can be expanded to other disease fields according to prior knowledge and actual requirements of doctors based on the constructed knowledge maps of the specific medical fields.
On the basis of the foregoing embodiment, in this embodiment, the first building block is specifically configured to: constructing a medical synonym library and a general concept knowledge map according to medical standards in the medical thesaurus; and constructing the concept knowledge graph of the target medical field according to the medical synonym library, the general concept knowledge graph and the prior knowledge of the doctor of the target medical field.
On the basis of the foregoing embodiment, in this embodiment, the first building module is further configured to: mapping the medical word stock from an ER data model to an RDF data model according to a mapping method which is selected by the doctor for the medical word stock in advance; and constructing the medical synonym library and the universal concept knowledge map according to the medical thesaurus of the RDF data model.
On the basis of the foregoing embodiment, in this embodiment, the first building block is further configured to: acquiring the concept and the relation input by the doctor, and representing the concept and the relation by a triple in an RDF data model; judging whether the input of the doctor exists in the medical word stock and the concept knowledge graph of the target medical field; if not, inserting the input into the medical thesaurus and the concept knowledge graph of the target medical field; if so, acquiring standard medical knowledge corresponding to the input according to the medical standard, and inserting the standard medical knowledge into the concept knowledge graph of the target medical field.
On the basis of the foregoing embodiment, the second building block in this embodiment is specifically configured to: according to a mapping method which is selected by the doctor for the electronic medical record in advance, mapping the structured data in the electronic medical record from an ER data model to an RDF data model; extracting entities, events and relations from unstructured data of the electronic medical record by adopting a machine learning or deep learning method according to the labels of doctors in the electronic medical record; and generating the instance knowledge graph according to the structured data of the EDF data model and the extracted entities, events and relations.
On the basis of the above embodiments, the mapping method in this embodiment is D2R, R2RML, or Virtuoso.
On the basis of the above embodiments, the medical word stock in this embodiment is a UMLS medical word stock.
Fig. 8 illustrates a physical structure diagram of an electronic device, and as shown in fig. 8, the electronic device may include: a processor (processor)801, a communication Interface (Communications Interface)802, a memory (memory)803 and a communication bus 804, wherein the processor 801, the communication Interface 802 and the memory 803 complete communication with each other through the communication bus 804. The processor 801 may call logic instructions in the memory 803 to perform the following method: constructing a concept knowledge graph of the target medical field according to medical standards in a medical word bank and prior knowledge of doctors in the target medical field; constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field; combining the concept knowledge graph and the instance knowledge graph into a fact knowledge graph.
In addition, the logic instructions in the memory 803 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The present embodiments provide a non-transitory computer-readable storage medium storing computer instructions that cause a computer to perform the methods provided by the above method embodiments, for example, including: constructing a concept knowledge graph of the target medical field according to medical standards in a medical word bank and prior knowledge of doctors in the target medical field; constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field; combining the concept knowledge graph and the instance knowledge graph into a fact knowledge graph.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A knowledge graph construction method aiming at a specific medical field is characterized by comprising the following steps:
constructing a concept knowledge graph of the target medical field according to medical standards in a medical word bank and prior knowledge of doctors in the target medical field;
constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field;
combining the concept knowledge graph and the instance knowledge graph into a fact knowledge graph.
2. The method of claim 1, wherein the step of constructing the conceptual knowledgebase of the target medical field based on the medical criteria in the medical thesaurus and the prior knowledge of the physician in the target medical field comprises:
constructing a medical synonym library and a general concept knowledge map according to medical standards in the medical thesaurus;
and constructing the concept knowledge graph of the target medical field according to the medical synonym library, the general concept knowledge graph and the prior knowledge of the doctor of the target medical field.
3. The method of claim 2, wherein the step of constructing the medical thesaurus and the generic concept knowledgebase according to medical criteria in the medical thesaurus comprises:
mapping the medical word stock from an entity relation model to a resource description framework model according to a mapping method preselected by the doctor for the medical word stock;
and constructing the medical synonym library and the general concept knowledge map according to the medical word library of the resource description framework model.
4. The method of claim 1, wherein the step of constructing the conceptual knowledgegraph of the target medical field based on the medical criteria in the medical thesaurus and the prior knowledge of the physician of the target medical field further comprises:
acquiring the concept and the relation input by the doctor, and representing the concept and the relation by a triple in a resource description framework model;
judging whether the input of the doctor exists in the medical word stock and the concept knowledge graph of the target medical field;
if not, inserting the input into the medical thesaurus and the concept knowledge graph of the target medical field;
if so, acquiring standard medical knowledge corresponding to the input according to the medical standard, and inserting the standard medical knowledge into the concept knowledge graph of the target medical field.
5. The method of claim 1, wherein the step of constructing an example knowledge graph of the target medical field according to an electronic medical record of the target medical field comprises:
according to a mapping method which is selected by the doctor for the electronic medical record in advance, mapping the structured data in the electronic medical record from an entity relation model to a resource description framework model;
extracting entities, events and relations from unstructured data of the electronic medical record by adopting a machine learning or deep learning method according to the labels of doctors in the electronic medical record;
and generating the instance knowledge graph according to the structured data of the resource description framework model and the extracted entities, events and relations.
6. The method of constructing a knowledge graph for specific medical fields according to claim 3 or 5, wherein the mapping method is D2R, R2RML or Virtuoso.
7. The method of constructing a knowledge graph for specific medical fields of claim 1, wherein the medical thesaurus is a UMLS medical thesaurus.
8. An apparatus for constructing a knowledge graph for a specific medical field, comprising:
the first construction module is used for constructing a concept knowledge graph of the target medical field according to the medical standard and the prior knowledge of doctors in the target medical field;
the second construction module is used for constructing an example knowledge graph of the target medical field according to the electronic medical record of the target medical field;
a combining module for combining the concept knowledge graph and the instance knowledge graph into a fact knowledge graph.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program performs the steps of the method of constructing a knowledge graph for a specific medical field according to any one of claims 1 to 7.
10. A non-transitory computer readable storage medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, implements the steps of the method for constructing a knowledge graph for a specific medical field according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911046367.6A CN110929040A (en) | 2019-10-30 | 2019-10-30 | Knowledge graph construction method and device for specific medical field |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911046367.6A CN110929040A (en) | 2019-10-30 | 2019-10-30 | Knowledge graph construction method and device for specific medical field |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110929040A true CN110929040A (en) | 2020-03-27 |
Family
ID=69850001
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911046367.6A Pending CN110929040A (en) | 2019-10-30 | 2019-10-30 | Knowledge graph construction method and device for specific medical field |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110929040A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111767410A (en) * | 2020-06-30 | 2020-10-13 | 平安国际智慧城市科技股份有限公司 | Construction method, device, equipment and storage medium of clinical medical knowledge map |
| CN111797243A (en) * | 2020-07-03 | 2020-10-20 | 中国烟草总公司湖南省公司 | Knowledge graph data system construction method, system, terminal and readable storage medium |
| CN111986799A (en) * | 2020-07-06 | 2020-11-24 | 北京欧应信息技术有限公司 | Orthopedics knowledge graph construction system taking joint movement function as core |
| CN112015905A (en) * | 2020-08-05 | 2020-12-01 | 河北工程大学 | Method for constructing fatigue marker disease knowledge graph |
| CN112463990A (en) * | 2020-12-17 | 2021-03-09 | 北京国电通网络技术有限公司 | Power grid infrastructure knowledge graph construction method and device, electronic equipment and storage medium |
| CN112966123A (en) * | 2021-03-02 | 2021-06-15 | 山东健康医疗大数据有限公司 | Medical health knowledge map system oriented to specific disease field |
| CN113539490A (en) * | 2021-06-10 | 2021-10-22 | 成都基预科技有限公司 | Common occupational disease risk prediction method based on knowledge graph |
| CN116383405A (en) * | 2023-03-20 | 2023-07-04 | 华中科技大学同济医学院附属协和医院 | Medical record knowledge graph construction method and system based on dynamic graph sequence |
-
2019
- 2019-10-30 CN CN201911046367.6A patent/CN110929040A/en active Pending
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111767410A (en) * | 2020-06-30 | 2020-10-13 | 平安国际智慧城市科技股份有限公司 | Construction method, device, equipment and storage medium of clinical medical knowledge map |
| CN111797243A (en) * | 2020-07-03 | 2020-10-20 | 中国烟草总公司湖南省公司 | Knowledge graph data system construction method, system, terminal and readable storage medium |
| CN111986799A (en) * | 2020-07-06 | 2020-11-24 | 北京欧应信息技术有限公司 | Orthopedics knowledge graph construction system taking joint movement function as core |
| CN112015905A (en) * | 2020-08-05 | 2020-12-01 | 河北工程大学 | Method for constructing fatigue marker disease knowledge graph |
| CN112463990A (en) * | 2020-12-17 | 2021-03-09 | 北京国电通网络技术有限公司 | Power grid infrastructure knowledge graph construction method and device, electronic equipment and storage medium |
| CN112966123A (en) * | 2021-03-02 | 2021-06-15 | 山东健康医疗大数据有限公司 | Medical health knowledge map system oriented to specific disease field |
| CN113539490A (en) * | 2021-06-10 | 2021-10-22 | 成都基预科技有限公司 | Common occupational disease risk prediction method based on knowledge graph |
| CN116383405A (en) * | 2023-03-20 | 2023-07-04 | 华中科技大学同济医学院附属协和医院 | Medical record knowledge graph construction method and system based on dynamic graph sequence |
| CN116383405B (en) * | 2023-03-20 | 2023-09-19 | 华中科技大学同济医学院附属协和医院 | Medical record knowledge graph construction method and system based on dynamic graph sequence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110929040A (en) | Knowledge graph construction method and device for specific medical field | |
| CN110990579B (en) | Cross-language medical knowledge graph construction method and device and electronic equipment | |
| Smith et al. | Relations in biomedical ontologies | |
| Beale | Archetypes: Constraint-based domain models for future-proof information systems | |
| US8589424B1 (en) | Method and system for accurate medical-code translation | |
| CN110427491B (en) | Medical knowledge graph construction method and device based on electronic medical record | |
| CA2747669C (en) | Method and system for validation of claims against policy with contextualized semantic interoperability | |
| Katayama et al. | The 3rd DBCLS BioHackathon: improving life science data integration with Semantic Web technologies | |
| CN110362690B (en) | Medical knowledge graph construction method and device | |
| Liu et al. | Natural language processing to ascertain two key variables from operative reports in ophthalmology | |
| Westra et al. | Modeling flowsheet data to support secondary use | |
| Agrawal et al. | Detecting modeling inconsistencies in SNOMED CT using a machine learning technique | |
| Kanterakis et al. | Towards reproducible bioinformatics: the OpenBio-C scientific workflow environment | |
| CN109299214B (en) | Text information extraction method, text information extraction device, text information extraction medium and electronic equipment | |
| Ceusters | The place of Referent Tracking in biomedical informatics | |
| Mohammed et al. | GraphQL patient case presentation using the problem oriented medical record schema | |
| Daniel et al. | Cross border semantic interoperability for learning health systems: the EHR4CR semantic resources and services | |
| Mo et al. | A decompositional approach to executing quality data model algorithms on the i2b2 platform | |
| CN111381744A (en) | Method, device and storage medium for establishing evaluation questionnaire | |
| Harman et al. | Standardized mapping of nursing assessments across 59 US military treatment facilities | |
| Chandrashekar et al. | Ontology mapping framework with feature extraction and semantic embeddings | |
| Pereira et al. | Querying semantic catalogues of biomedical databases | |
| Gazzarata et al. | A standardized SOA for clinical data sharing to support acute care, telemedicine and clinical trials | |
| Barcellos et al. | A reliable approach for applying DICOM structured reporting in a large-scale telemedicine network | |
| Juhrisch et al. | Information systems engineering in healthcare–an evaluation of the state of the art of operational process design |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200717 Address after: 100084 Beijing Haidian District Tsinghua Yuan 100084-82 mailbox Applicant after: TSINGHUA University Applicant after: BEIJING TSINGHUA CHANGGUNG Hospital Address before: 100084 Beijing Haidian District Tsinghua Yuan 100084-82 mailbox Applicant before: TSINGHUA University |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200327 |