CN111599479A - ICD 9-CM-3-based surgical knowledge map construction method and device - Google Patents
ICD 9-CM-3-based surgical knowledge map construction method and device Download PDFInfo
- Publication number
- CN111599479A CN111599479A CN202010256527.6A CN202010256527A CN111599479A CN 111599479 A CN111599479 A CN 111599479A CN 202010256527 A CN202010256527 A CN 202010256527A CN 111599479 A CN111599479 A CN 111599479A
- Authority
- CN
- China
- Prior art keywords
- concept
- concepts
- codes
- code
- relation
- 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
Images
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
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides an ICD 9-CM-3-based operation knowledge graph construction method and device, and the method comprises the following steps: step 1, fusing all the ICD-9-CM-3 official and national extended versions of operation concepts to obtain fusion information and establish a guide code of the relationship between the operation concepts; step 2, constructing a relation between operation concepts based on the guide codes, and if the codes corresponding to the operation concepts have implication relations, constructing the implication relations between the operations according to the implication relations of the codes; step 3, defining the concept attribute of the operation; and 4, step 4: carrying out named entity identification on the operation concept by using a Bert + BilSTM + CRF model, and identifying defined concept attributes; and 5: extracting the identified concept attributes, and establishing attribute relations between the concept attributes and the corresponding operation concepts; step 6: and filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
Description
Technical Field
The invention provides an ICD 9-CM-3-based operation knowledge graph construction method and device, and belongs to the technical field of medical knowledge graph construction.
Background
One construction method of the existing medical knowledge graph is to extract medical entities (such as diseases, symptoms and examinations) from medical data (such as medical records) by using a natural language processing technology, then to construct the correlation between the medical entities by calculating the correlation between the entities, and finally to import the medical entities and the correlation into a graph database. The traditional machine learning method is adopted to mine medical entities and relationships from medical records, and no attempt is made to extract and fuse medical entities and relationships from an official and authorized operation term set. There is no reference to the definition and extraction of entity attributes.
Disclosure of Invention
The invention provides an ICD 9-CM-3-based operation knowledge graph construction method and device, which are used for solving the problem of low accuracy of a graph constructed by the existing method, and adopt the following technical scheme:
an ICD 9-CM-3-based surgical knowledge map construction method, wherein the surgical knowledge map construction method comprises the following steps: firstly, constructing a guide code of operation concept relation by using ICD-9-CM-3 official and all national extension operation concepts, and constructing the relation and implication relation between the operation concepts by using the guide code; then, defining the concept attribute of the operation according to the operation concept and the attribute value, and identifying the operation concept by utilizing a Bert + BilSTM + CRF model to obtain the concept attribute; and finally, establishing an attribute relationship between the concept attribute and the operation concept, and forming an operation knowledge graph by using the attribute relationship and the implication relationship.
Further, the specific process of the operation knowledge map construction method comprises the following steps:
step 1, fusing ICD-9-CM-3 official and all national extended versions of operation concepts to obtain fusion information containing the ICD-9-CM-3 official and all national extended versions of operation concepts, and constructing a guide code of the relationship between the operation concepts according to the fusion information;
step 2, constructing a relation between operation concepts based on the guide codes obtained in the step 1, and if the codes corresponding to the operation concepts have implication relations, constructing implication relations between operations according to the coding implication relations;
step 3, defining the concept attribute of the operation according to the operation concept and the attribute value;
and 4, step 4: carrying out named entity identification on the operation concept by using a Bert + BilSTM + CRF model, and identifying the concept attribute defined in the step 3;
and 5: extracting the concept attributes identified in the step 4, and establishing an attribute relationship between the concept attributes and the operation concepts corresponding to the concept attributes;
step 6: and filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
Further, the step 1 of constructing the guide code comprises:
step 11, judging whether all the coding formats of a plurality of operation concepts exist in the operation concepts in the fusion information are extension codes, if not, extracting the coarse-grained codes of the operation concepts as guide codes aiming at the operation concepts with the same coarse granularity in the coding formats in the fusion information; if yes, jumping to step 12;
step 12, aiming at a plurality of operation concepts of which the encoding formats are all extension encoding, selecting one of the operation concepts as a main version, and taking the encoding corresponding to the operation concept as the main version as guide encoding;
and step 13, judging whether the expanded codes of the operation concepts except the operation concepts of the main version contain codes and concepts which do not exist in the codes of the main version, if so, cutting all the expanded codes of the versions corresponding to the operation concepts which contain the codes which do not exist in the codes of the main version into codes of the previous level, voting according to the times of the appearance of the codes of the previous level, and selecting the expanded code corresponding to the most code as the guide code.
Furthermore, a self-supervision method is adopted to represent the word learning characteristics on the basis of the Bert medical corpus in the Bert + BilSTM + CRF model, and the CRF restricts the orderliness of the BilSTM output labels through transfer characteristics.
A surgical knowledge map constructing apparatus corresponding to any one of the above methods, the surgical knowledge map constructing apparatus comprising:
the information fusion module is used for fusing ICD-9-CM-3 official concepts and all national extended version operation concepts to obtain fusion information containing the ICD-9-CM-3 official concepts and all national extended version operation concepts;
the relation and coding construction module block is used for constructing guide codes of the relation between the operation concepts, and the relation and implication relation between the operation concepts;
the concept attribute definition module is used for defining the concept attribute of the operation according to the operation concept and the attribute value;
the Bert + BiLSTM + CRF model is used for carrying out named entity identification on the operation concept and identifying the concept attribute corresponding to the operation concept;
the attribute relationship establishing module is used for extracting the concept attributes identified by the Bert + BiLSTM + CRF model and establishing attribute relationships between the concept attributes and the corresponding operation concepts;
and the import module is used for filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
Further, the relationship and coding building block comprises:
the guiding code establishing module is used for establishing guiding codes of the relation between the operation concepts according to the fusion information;
the relation establishing module is used for establishing the relation between the operation concepts according to the guide codes, judging whether the implication relation exists between the codes corresponding to the operation concepts or not, and if the implication relation exists between the codes corresponding to the operation concepts, establishing the implication relation between the operations according to the code implication relation;
further, the guide code establishing module comprises:
the coding format judging module is used for judging whether all the coding formats of a plurality of operation concepts in the fusion information are extended codes, and if not, the coarse-grained guide coding establishing module is started; if yes, starting a main version mode to guide a code establishing module;
the coarse-granularity mode guidance code establishing module is used for extracting coarse-granularity codes of the operation concepts as guidance codes aiming at the operation concepts with the same coarse granularity in the coding formats in the fusion information;
the main version mode guiding code establishing module is used for selecting one of a plurality of operation concepts as a main version and selecting a code corresponding to the operation concept as the main version as a guiding code aiming at the plurality of operation concepts of which all encoding formats are extension codes;
the system comprises a main version of operation concept, a spread code mode guiding code establishing module, a code searching module and a code searching module, wherein the main version of operation concept comprises a main version of operation concept, the main version of operation concept comprises a main version of.
Furthermore, a self-supervision method is adopted to represent the word learning characteristics on the basis of the Bert medical corpus in the Bert + BilSTM + CRF model, and the CRF restricts the orderliness of the BilSTM output labels through transfer characteristics.
The invention has the beneficial effects that:
the method and the device for constructing the operation knowledge graph are used for establishing the operation knowledge graph based on authoritative ICD-9-CM-3 official authorities and extended versions all over the country, and the accuracy of the operation concept and relationship of the operation knowledge graph constructed by the method is higher.
Drawings
FIG. 1 is a flow chart of a surgical knowledge map construction method of the invention;
FIG. 2 is a schematic structural diagram of the surgical knowledge map constructing device of the invention;
FIG. 3 is a schematic diagram of the surgical knowledge map constructing apparatus according to the present invention.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
The embodiment of the invention provides an ICD 9-CM-3-based operation knowledge graph construction method and device, and aims to solve the problem that an accuracy rate of a graph constructed by an existing method is low.
An ICD 9-CM-3-based surgical knowledge map construction method, wherein the surgical knowledge map construction method comprises the following steps: firstly, constructing a guide code of operation concept relation by using ICD-9-CM-3 official and all national extension operation concepts, and constructing the relation and implication relation between the operation concepts by using the guide code; then, defining the concept attribute of the operation according to the operation concept and the attribute value, and identifying the operation concept by utilizing a Bert + BilSTM + CRF model to obtain the concept attribute; and finally, establishing an attribute relationship between the concept attribute and the operation concept, and forming an operation knowledge graph by using the attribute relationship and the implication relationship.
The working principle and the technical effect of the technical scheme are as follows: the operation knowledge graph is established based on authoritative ICD-9-CM-3 official authorities and expansion versions all over the country, and the accuracy of the operation concept and relationship of the operation knowledge graph established by the method is higher.
In one embodiment of the present invention, as shown in fig. 1, the method for constructing a surgical knowledge map includes:
step 1, fusing ICD-9-CM-3 official and all national extended versions of operation concepts to obtain fusion information containing the ICD-9-CM-3 official and all national extended versions of operation concepts, and constructing a guide code of the relationship between the operation concepts according to the fusion information; wherein, the surgery concept refers to the name of the surgery;
step 2, constructing a relation between operation concepts based on the guide codes obtained in the step 1, and if the codes corresponding to the operation concepts have implication relations, constructing implication relations between operations according to the coding implication relations;
step 3, determining the concept attribute of the operation by using a Schmea definition method according to the concept and the attribute value of the operation; wherein the concept attribute is attribute information related in the operation corresponding to the operation name;
and 4, step 4: carrying out named entity identification on the operation concept by using a Bert + BilSTM + CRF model, and identifying the concept attribute defined in the step 3;
and 5: extracting the concept attributes identified in the step 4, and establishing an attribute relationship between the concept attributes and the operation concepts corresponding to the concept attributes;
step 6: and filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
The working principle and the beneficial effects of the scheme are as follows: the ICD-9-CM-3 official and all the national extension operating concepts are fused into a large information base, and a guide code of the relationship between the operating concepts is constructed according to the fusion information; then, establishing a relation between the operation concepts according to the guide codes, and if the codes corresponding to the operation concepts have implication relations, establishing implication relations between the operations according to the coding implication relations; for example, for two surgical concepts: 07.22 unilateral adrenal resection and 07.2201 laparoscopic unilateral adrenal resection, since 07.2201 contains 07.22, the laparoscopic unilateral adrenal resection is a sub-concept of the unilateral adrenal resection. The concept attributes of the surgery are defined according to the surgery concept and the attribute values, and the concept attributes are defined as shown in table 1. Carrying out named entity identification on the operation concept by using a Bert + BilSTM + CRF model, and identifying the concept attribute defined in the step 3; extracting the identified concept attributes, and establishing attribute relations between the concept attributes and the corresponding operation concepts; and finally, filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
According to the operation knowledge graph construction method based on ICD9-CM-3, the operation knowledge graph is established based on authoritative ICD-9-CM-3 official and nationwide extended versions, and the accuracy of operation concepts and relations of the operation knowledge graph constructed by the method is higher.
TABLE 1
| Concept attributes | Attribute value | Examples of such applications are |
| Location of a body part | Array of elements | 'sternum' |
| Surgical formula | Array of elements | [ "removal technique"] |
| Abnormal form | Character string | Ulcer " |
| Access road | Array of elements | [ "nostril approach"] |
| Special implements | Character string | Electric heating ring " |
| Manipulation of | Character string | "Clamp" or vice " |
| Endoscope | Character string | Cystoscope " |
| Range of | Character string | "local" |
| Orientation | Array of elements | [ "left side"] |
| Purpose(s) to | Character string | [ "terminating pregnancy"] |
| Operation of | Character string | "Kidney dialysis" |
In an embodiment of the present invention, the step 1 of constructing the guide code includes:
step 11, judging whether all the coding formats of a plurality of operation concepts exist in the operation concepts in the fusion information are extension codes, if not, extracting the coarse-grained codes of the operation concepts as guide codes aiming at the operation concepts with the same coarse granularity in the coding formats in the fusion information; if yes, jumping to step 12;
step 12, aiming at a plurality of operation concepts of which the encoding formats are all extension encoding, selecting one of the operation concepts as a main version, and taking the encoding corresponding to the operation concept as the main version as guide encoding; other surgical concepts than the surgical concept as the main version are other versions;
and step 13, judging whether the expanded codes of the operation concepts except the operation concepts of the main version contain codes and concepts which do not exist in the codes of the main version, if so, cutting all the expanded codes of the versions corresponding to the operation concepts which contain the codes which do not exist in the codes of the main version into codes of the previous level, voting according to the times of the appearance of the codes of the previous level, and selecting the expanded code corresponding to the most code as the guide code.
The working principle and the beneficial effects of the scheme are as follows: the present embodiment constructs the guide code in three ways:
firstly, constructing guide codes according to a coarse-grained mode, judging whether all the coding formats of a plurality of operation concepts in the fusion information are extension codes, and if not, extracting the coarse-grained codes of the operation concepts as the guide codes aiming at the operation concepts with the same coarse granularity in the coding formats in the fusion information; for example: an extraocular muscle lengthening operation 15.21 international disease classification surgery code ICD-9-CM-3, 15.2100 operation classification and code (2018) of Shandong medical institution, 15.2100 standard library of Shanghai operation classification and code (2018), 15.2100 clinical edition of operation classification and code country 2.0, 15.21 is selected as a guide code.
And if so, constructing the guide code in a main version mode.
Then, constructing a guidance code by adopting a main version mode, and selecting one of a plurality of operation concepts as a main version and a code corresponding to the operation concept as the main version as the guidance code aiming at the plurality of operation concepts of which all the encoding formats are extension codes; other surgical concepts than the one that is the main version are other versions. For example, if all of the plurality of codes are extension codes, a certain extension version is selected as the master version, and a code corresponding to the version is selected as the guidance code, so that "surgical procedure classification and code national clinical version 2.0" can be used as the master version.
And finally, constructing guide codes in a spreading code mode, judging whether spread codes of the operation concepts except the operation concepts of the main version contain codes and concepts which do not exist in the codes of the main version or not, if so, cutting all the version spread codes corresponding to the operation concepts which contain the codes which do not exist in the codes of the main version into codes of the previous level, voting according to the times of the appearance of the codes of the previous level, and selecting the spread code corresponding to the most code as the guide code. Wherein, the upper level coding refers to the level one coding with the processing level prior to the extension coding.
The guidance code is extracted through various modes, the guidance code which accords with the content of the hand surgery concept in the fusion information can be more accurately constructed according to the code corresponding to the surgery concept, the matching degree between the guidance code and the surgery concept is improved, and the accuracy between the surgery concept and the relation of the surgery knowledge graph is further improved.
According to one embodiment of the invention, the Bert in the Bert + BilSTM + CRF model can be used for learning words to be represented by characteristics which are more accurate and contain more information by adopting a self-supervision (Schmea) method on the basis of mass medical linguistic data. BilSTM can better capture bidirectional semantic dependency in surgical concepts. The CRF constrains the sequentiality of the BiLSTM output labels by transferring features.
The working principle and the beneficial effects of the scheme are as follows: carrying out named entity identification on the operation concept through a Bert + BilSt + CRF model and a Schmea extraction method, and identifying the concept attribute corresponding to the operation concept; the accuracy and the richness of feature representation in the map construction process can be effectively improved, the bidirectional semantic dependency relationship in the operation concept can be captured more accurately, the sequence of the BilSTM output label can be restrained by transferring features in the past eight years, and the disorder phenomenon cannot occur in the output process.
The embodiment of the invention provides an operation knowledge graph construction device corresponding to any one of the methods, and as shown in fig. 2, the device comprises an information fusion module, a relation and coding construction module block, a concept attribute definition module, a Bert + BilSTM + CRF model, an attribute relation establishment module and an import module;
the information fusion module is used for fusing ICD-9-CM-3 official concepts and all national extended version operation concepts to obtain fusion information containing the ICD-9-CM-3 official concepts and all national extended version operation concepts;
the relation and coding construction module block is used for constructing guide codes of the relation between the operation concepts, and the relation and implication relation between the operation concepts;
the concept attribute definition module is used for defining the concept attribute of the operation according to the operation concept and the attribute value;
the Bert + BiLSTM + CRF model is used for carrying out named entity identification on the operation concept and identifying the concept attribute corresponding to the operation concept;
the attribute relationship establishing module is used for extracting the concept attributes identified by the Bert + BiLSTM + CRF model and establishing attribute relationships between the concept attributes and the corresponding operation concepts;
and the import module is used for filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
The working principle and the beneficial effects of the scheme are as follows: fusing ICD-9-CM-3 official and all national expanded version operation concepts to obtain fusion information containing the ICD-9-CM-3 official and all national expanded version operation concepts, and constructing a guide code of the relationship between the operation concepts according to the fusion information; then, constructing a relation between operation concepts through the obtained guide codes, and if the codes corresponding to the operation concepts have implication relations, constructing implication relations between operations according to the coding implication relations; determining the concept attribute of the operation by using a Schmea definition method according to the concept and attribute value of the operation; carrying out named entity identification on the operation concept by using a Bert + BilSTM + CRF model, and identifying defined concept attributes; finally, extracting the identified concept attributes, and establishing attribute relations between the concept attributes and the corresponding operation concepts; and filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
According to the operation knowledge graph construction method based on ICD9-CM-3, the operation knowledge graph is established based on authoritative ICD-9-CM-3 official and nationwide extended versions, and the accuracy of operation concepts and relations of the operation knowledge graph constructed by the method is higher.
In one embodiment of the present invention, the relationship and coding building block includes:
the guiding code establishing module is used for establishing guiding codes of the relation between the operation concepts according to the fusion information;
the relation establishing module is used for establishing the relation between the operation concepts according to the guide codes, judging whether the implication relation exists between the codes corresponding to the operation concepts or not, and if the implication relation exists between the codes corresponding to the operation concepts, establishing the implication relation between the operations according to the code implication relation;
the working principle and the technical effect of the technical scheme are as follows: and constructing guide codes of the relation between the operation concepts according to the fusion information, constructing the relation between the operation concepts according to the guide codes, judging whether the implication relation exists between the codes corresponding to the operation concepts, and constructing the implication relation between the operations according to the code implication relation if the implication relation exists between the codes corresponding to the operation concepts. The accuracy and the efficiency of acquiring the guide codes and the implication relation are effectively improved, and the accuracy and the efficiency of map construction are further improved.
In an embodiment of the present invention, the guide code creating module includes:
the coding format judging module is used for judging whether all the coding formats of a plurality of operation concepts in the fusion information are extended codes, and if not, the coarse-grained guide coding establishing module is started; if yes, starting a main version mode to guide a code establishing module;
the coarse-granularity mode guidance code establishing module is used for extracting coarse-granularity codes of the operation concepts as guidance codes aiming at the operation concepts with the same coarse granularity in the coding formats in the fusion information;
the main version mode guiding code establishing module is used for selecting one of a plurality of operation concepts as a main version and selecting a code corresponding to the operation concept as the main version as a guiding code aiming at the plurality of operation concepts of which all encoding formats are extension codes;
the system comprises a main version of operation concept, a spread code mode guiding code establishing module, a code searching module and a code searching module, wherein the main version of operation concept comprises a main version of operation concept, the main version of operation concept comprises a main version of.
The working principle and the beneficial effects of the scheme are as follows: the method is characterized in that a coarse-grained mode guidance code establishing module, a major version mode guidance code establishing module and a spreading code mode guidance code establishing module are used for guiding the construction of codes according to a coding format corresponding to a surgical concept by adopting different construction modes.
The guidance code is extracted through various modes, the guidance code which accords with the content of the fusion information operation concept can be more accurately constructed according to the code corresponding to the operation concept, the matching degree between the guidance code and the operation concept is improved, and the accuracy between the operation concept and the relation of the operation knowledge graph is further improved.
According to one embodiment of the invention, a self-supervision method is adopted as word learning characteristic representation on the basis of the Bert + BilSTM + CRF medical corpus in the model, and the CRF restrains the sequentiality of the BilSTM output labels through transfer characteristics.
The working principle and the beneficial effects of the scheme are as follows: the Bert in the Bert + BilSTM + CRF model can learn words to be represented by more accurate characteristics containing more abundant information by adopting a self-supervision (Schmea) method on the basis of massive medical linguistic data. BilSTM can better capture bidirectional semantic dependency in surgical concepts. The CRF constrains the sequentiality of the BiLSTM output labels by transferring features. Carrying out named entity identification on the operation concept through a Bert + BilSt + CRF model and a Schmea extraction method, and identifying the concept attribute corresponding to the operation concept; the accuracy and the richness of feature representation in the map construction process can be effectively improved, the bidirectional semantic dependency relationship in the operation concept can be captured more accurately, the sequence of the BilSTM output label can be restrained by transferring features in the past eight years, and the disorder phenomenon cannot occur in the output process.
ICD9-CM-3 is a coding system for surgery and operation with a hierarchical relationship. It comprises the following components:
chapter 06-07 surgery on the endocrine System
Category 07 other endocrine gland surgeries
Sub-mesh 07.2 partial adrenal gland resection
Unilateral adrenal resection of fine mesh 07.22
Extension code 07.2200 unilateral adrenal resection (operation classification code national clinic version 2.0)
Extension code 07.2201 laparoscopic unilateral adrenal resection (surgical procedures classification code national clinic version 2.0)
Extension code 07.2200y001 unilateral adrenal resection (national clinical 1.1 version operation and operation (ICD-9-CM-3)2018 extension (Yunnan province))
All the expansion versions nationwide are ICD-9-CM-3 expansion versions, and at least comprise:
ICD-9-CM-3 list 2011 edition (Liu love people main edition)
National clinical edition of surgical operation classification code 1.1
Surgical procedure Classification and code national version 2017
Version V6.01 of operation name and code standard of first page of admission case in Beijing
Guangdong province ICD-9-CM-3 operation and operation code (2017 edition)
Guangdong province ICD-9-CM-3 operation and operation code (2016 type)
Shandong province medical institution operation classification code and level catalog (2018)
Shandong province-clinical edition international disease classification ICD-9-CM-3(V6.01 edition)
Beijing department inpatient case first page operation name and code standard v5.0
Beijing edition operation name v6.0
Beijing RC022-ICD-9 surgical coding
ICD operation code of Sichuan province
Operation code ICD-9-CM-3(2017 maintenance type)
National Standard 1.0-surgical operation Classification code national clinical edition 1.0
Group Standard of TCHIA001-2017 operation, operation Classification and code
National clinical edition of surgical operation classification code 2.0
Standard library of operation and codes for 2018 years in Shanghai
National clinic 1.1 edition operation and operation (ICD-9-CM-3)2018 extension (Yunnan province)
Surgical operation classification and code (ICD-9-CM-3)2019 medical insurance edition
ICD9-2017 harmonization clinical edition
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (8)
1. An ICD 9-CM-3-based surgical knowledge map construction method is characterized in that the surgical knowledge map construction method comprises the following steps: firstly, constructing a guide code of operation concept relation by using ICD-9-CM-3 official and all national extension operation concepts, and constructing the relation and implication relation between the operation concepts by using the guide code; then, defining the concept attribute of the operation according to the operation concept and the attribute value, and identifying the operation concept by utilizing a Bert + BilSTM + CRF model to obtain the concept attribute; and finally, establishing an attribute relationship between the concept attribute and the operation concept, and forming an operation knowledge graph by using the attribute relationship and the implication relationship.
2. The surgical knowledge graph construction method according to claim 1, wherein the specific process of the surgical knowledge graph construction method comprises:
step 1, fusing ICD-9-CM-3 official and all national extended versions of operation concepts to obtain fusion information containing the ICD-9-CM-3 official and all national extended versions of operation concepts, and constructing a guide code of the relationship between the operation concepts according to the fusion information;
step 2, constructing a relation between operation concepts based on the guide codes obtained in the step 1, and if the codes corresponding to the operation concepts have implication relations, constructing implication relations between operations according to the coding implication relations;
step 3, defining the concept attribute of the operation according to the operation concept and the attribute value;
and 4, step 4: carrying out named entity identification on the operation concept by using a Bert + BilSTM + CRF model, and identifying the concept attribute defined in the step 3;
and 5: extracting the concept attributes identified in the step 4, and establishing an attribute relationship between the concept attributes and the operation concepts corresponding to the concept attributes;
step 6: and filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
3. The surgical knowledge map construction method according to claim 1, wherein the step 1 of constructing the guide code comprises:
step 11, judging whether all the coding formats of a plurality of operation concepts exist in the operation concepts in the fusion information are extension codes, if not, extracting the coarse-grained codes of the operation concepts as guide codes aiming at the operation concepts with the same coarse granularity in the coding formats in the fusion information; if yes, jumping to step 12;
step 12, aiming at a plurality of operation concepts of which the encoding formats are all extension encoding, selecting one of the operation concepts as a main version, and taking the encoding corresponding to the operation concept as the main version as guide encoding;
and step 13, judging whether the expanded codes of the operation concepts except the operation concepts of the main version contain codes and concepts which do not exist in the codes of the main version, if so, cutting all the expanded codes of the versions corresponding to the operation concepts which contain the codes which do not exist in the codes of the main version into codes of the previous level, voting according to the times of the appearance of the codes of the previous level, and selecting the expanded code corresponding to the most code as the guide code.
4. The method for constructing an operative knowledge graph according to claim 1, wherein a self-supervision method is adopted as a word learning feature representation on the basis of the Bert medical corpus in the Bert + BilSTM + CRF model, and the CRF restrains the sequentiality of the BilSTM output labels through transfer features.
5. A surgical knowledge map construction apparatus according to any one of claims 1 to 4, wherein the apparatus comprises
The information fusion module is used for fusing ICD-9-CM-3 official concepts and all national extended version operation concepts to obtain fusion information containing the ICD-9-CM-3 official concepts and all national extended version operation concepts;
the relation and coding construction module block is used for constructing guide codes of the relation between the operation concepts, and the relation and implication relation between the operation concepts;
the concept attribute definition module is used for defining the concept attribute of the operation according to the operation concept and the attribute value;
the Bert + BiLSTM + CRF model is used for carrying out named entity identification on the operation concept and identifying the concept attribute corresponding to the operation concept;
the attribute relationship establishing module is used for extracting the concept attributes identified by the Bert + BiLSTM + CRF model and establishing attribute relationships between the concept attributes and the corresponding operation concepts;
and the import module is used for filling the attribute relation and the implication relation into a graph database to form a surgical knowledge graph.
6. The surgical knowledge map construction apparatus of claim 5 wherein the relationship and coding construction module block comprises:
the guiding code establishing module is used for establishing guiding codes of the relation between the operation concepts according to the fusion information;
and the relation establishing module is used for establishing the relation between the operation concepts according to the guide codes, judging whether the implication relation exists between the codes corresponding to the operation concepts or not, and if the implication relation exists between the codes corresponding to the operation concepts, establishing the implication relation between the operations according to the code implication relation.
7. The surgical knowledge map construction apparatus of claim 6, wherein the guide code creation module comprises:
the coding format judging module is used for judging whether all the coding formats of a plurality of operation concepts in the fusion information are extended codes, and if not, the coarse-grained guide coding establishing module is started; if yes, starting a main version mode to guide a code establishing module;
the coarse-granularity mode guidance code establishing module is used for extracting coarse-granularity codes of the operation concepts as guidance codes aiming at the operation concepts with the same coarse granularity in the coding formats in the fusion information;
the main version mode guiding code establishing module is used for selecting one of a plurality of operation concepts as a main version and selecting a code corresponding to the operation concept as the main version as a guiding code aiming at the plurality of operation concepts of which all encoding formats are extension codes;
the system comprises a main version of operation concept, a spread code mode guiding code establishing module, a code searching module and a code searching module, wherein the main version of operation concept comprises a main version of operation concept, the main version of operation concept comprises a main version of.
8. The surgical knowledge graph constructing apparatus according to claim 5, wherein the Bert + BilSTM + CRF model employs a self-supervision method as a word learning feature representation on the basis of the Bert medical corpus, and the CRF constrains the orderliness of the BilSTM output labels by transferring features.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010256527.6A CN111599479B (en) | 2020-04-02 | 2020-04-02 | ICD 9-CM-3-based surgical knowledge graph construction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010256527.6A CN111599479B (en) | 2020-04-02 | 2020-04-02 | ICD 9-CM-3-based surgical knowledge graph construction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111599479A true CN111599479A (en) | 2020-08-28 |
| CN111599479B CN111599479B (en) | 2023-08-11 |
Family
ID=72181886
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010256527.6A Active CN111599479B (en) | 2020-04-02 | 2020-04-02 | ICD 9-CM-3-based surgical knowledge graph construction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111599479B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113204968A (en) * | 2021-05-28 | 2021-08-03 | 平安科技（深圳）有限公司 | Concept recognition method, device, equipment and storage medium of medical entity |
| CN113779179A (en) * | 2021-09-29 | 2021-12-10 | 北京雅丁信息技术有限公司 | ICD intelligent coding method based on deep learning and knowledge graph |
| WO2022252107A1 (en) * | 2021-06-01 | 2022-12-08 | 眼灵(上海)智能科技有限公司 | Disease examination system and method based on eye image |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140257842A1 (en) * | 2006-03-27 | 2014-09-11 | A-Life Medical, Llc | Auditing the coding and abstracting of documents |
| US20160132648A1 (en) * | 2014-11-06 | 2016-05-12 | ezDI, LLC | Data Processing System and Method for Computer-Assisted Coding of Natural Language Medical Text |
| CN107145744A (en) * | 2017-05-08 | 2017-09-08 | 合肥工业大学 | Construction method, device and the aided diagnosis method of medical knowledge collection of illustrative plates |
| CN107423820A (en) * | 2016-05-24 | 2017-12-01 | 清华大学 | The knowledge mapping of binding entity stratigraphic classification represents learning method |
| CN109522552A (en) * | 2018-11-09 | 2019-03-26 | 天津开心生活科技有限公司 | A kind of method for normalizing of medical information, device, medium and electronic equipment |
| CN110781254A (en) * | 2020-01-02 | 2020-02-11 | 四川大学 | Automatic case knowledge graph construction method, system, equipment and medium |
-
2020
- 2020-04-02 CN CN202010256527.6A patent/CN111599479B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140257842A1 (en) * | 2006-03-27 | 2014-09-11 | A-Life Medical, Llc | Auditing the coding and abstracting of documents |
| US20160132648A1 (en) * | 2014-11-06 | 2016-05-12 | ezDI, LLC | Data Processing System and Method for Computer-Assisted Coding of Natural Language Medical Text |
| CN107423820A (en) * | 2016-05-24 | 2017-12-01 | 清华大学 | The knowledge mapping of binding entity stratigraphic classification represents learning method |
| CN107145744A (en) * | 2017-05-08 | 2017-09-08 | 合肥工业大学 | Construction method, device and the aided diagnosis method of medical knowledge collection of illustrative plates |
| CN109522552A (en) * | 2018-11-09 | 2019-03-26 | 天津开心生活科技有限公司 | A kind of method for normalizing of medical information, device, medium and electronic equipment |
| CN110781254A (en) * | 2020-01-02 | 2020-02-11 | 四川大学 | Automatic case knowledge graph construction method, system, equipment and medium |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113204968A (en) * | 2021-05-28 | 2021-08-03 | 平安科技（深圳）有限公司 | Concept recognition method, device, equipment and storage medium of medical entity |
| WO2022252107A1 (en) * | 2021-06-01 | 2022-12-08 | 眼灵(上海)智能科技有限公司 | Disease examination system and method based on eye image |
| CN113779179A (en) * | 2021-09-29 | 2021-12-10 | 北京雅丁信息技术有限公司 | ICD intelligent coding method based on deep learning and knowledge graph |
| CN113779179B (en) * | 2021-09-29 | 2024-02-09 | 北京雅丁信息技术有限公司 | ICD intelligent coding method based on deep learning and knowledge graph |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111599479B (en) | 2023-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111599479A (en) | ICD 9-CM-3-based surgical knowledge map construction method and device | |
| CN109492077A (en) | The petrochemical field answering method and system of knowledge based map | |
| CN106776711A (en) | A kind of Chinese medical knowledge mapping construction method based on deep learning | |
| CN111708899B (en) | Engineering information intelligent searching method based on natural language and knowledge graph | |
| CN109308321A (en) | A kind of knowledge question answering method, knowledge Q-A system and computer readable storage medium | |
| CN115292457B (en) | Knowledge question answering method and device, computer readable medium and electronic equipment | |
| CN104899340B (en) | A kind of IETM technical information fragment retrieval device and its search method based on fragment of most compacting | |
| CN104866593A (en) | Database searching method based on knowledge graph | |
| CN103488724A (en) | Book-oriented reading field knowledge map construction method | |
| CN111309930B (en) | Medical knowledge graph entity alignment method based on representation learning | |
| CN110175334A (en) | Text knowledge's extraction system and method based on customized knowledge slot structure | |
| CN110188359B (en) | Text entity extraction method | |
| CN112749265A (en) | Intelligent question-answering system based on multiple information sources | |
| CN112559770A (en) | Text data relation extraction method, device and equipment and readable storage medium | |
| CN115293161A (en) | Reasonable medicine taking system and method based on natural language processing and medicine knowledge graph | |
| CN112966117A (en) | Entity linking method | |
| CN111984694A (en) | Orthopedics search engine system | |
| CN114334105A (en) | Named entity identification-based disease transcoding method and device | |
| CN112989811B (en) | History book reading auxiliary system based on BiLSTM-CRF and control method thereof | |
| US20080015843A1 (en) | Linguistic Image Label Incorporating Decision Relevant Perceptual, Semantic, and Relationships Data | |
| CN115270774B (en) | Big data keyword dictionary construction method for semi-supervised learning | |
| CN116775826A (en) | Intelligent question-answering method and device, electronic equipment and storage medium | |
| CN101246473B (en) | Segmentation system evaluating method and segmentation evaluating system | |
| CN116168844A (en) | Medical data processing system based on big data analysis | |
| CN113468307A (en) | Text processing method and device, electronic equipment and storage medium |
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 |