Knowledge graph-based medical terminology mapping methods, electronic devices, and media

Abstract

This application discloses a knowledge graph-based method, electronic device, and medium for mapping medical terms. The method may include: determining the basic architecture type of the knowledge graph for medical terms, then assigning relevance values ​​to obtain the knowledge graph; performing format standardization processing on the set of terms to be mapped, matching it with the knowledge graph, and obtaining the mapping result for the mapped terms; after obtaining the mapping result, performing semi-automatic mapping based on the knowledge graph for new terms in the set of terms to obtain the mapping result corresponding to the new terms. This invention combines knowledge graphs with medical terminology mapping, increasing the relevance of term attributes, improving the compatibility of terminology mapping, and automatically correcting the mapping effect of newly added terms based on the knowledge graph.

Description

Technical Field

[0001] This invention relates to the field of medical terminology mapping, and more specifically, to a knowledge graph-based method, electronic device, and medium for medical terminology mapping. Background Technology

[0002] Medical terminology refers to the specialized vocabulary used by healthcare professionals in diagnosis, treatment, and care to achieve effective and accurate communication. It is a crucial component of clinical information systems, encompassing diagnostic terminology, pharmaceutical terminology, laboratory terminology, examination terminology, procedural data, surgical data, and traditional Chinese medicine terminology. Internationally, medical terminology is more standardized, including the International Classification of Diseases (ICD-10), the International Classification of Diseases (ICD-11), the International Classification of Diseases, Ninth Edition, Clinical Revision, Volume 3 (ICD-9-CM-3), the Standard Nomenclature for Clinical Drugs (RxNorm), Systematic Nomenclature of Medical Terminology - Clinical Terminology (SNOMED-CT), the Logical Nomenclature Precoding System for Observational Indicators (LOINC), and the Unified Medical Language System (UMLS). Domestically, there are also standards such as "Common Clinical Medical Terminology (2023 Edition)," the National Clinical Version 2.0 Disease Diagnosis Code (ICD-10), and the National Medical Insurance Version 2.0 Disease Diagnosis Code (ICD-10).

[0003] However, domestic medical terminology standards are lacking and the system is incomplete. Most standards are designed for statistical and billing convenience and differ from clinical terminology. This is mainly manifested in the following ways: 1. Lack of a unified organizational structure; terminology is maintained independently by various health administrative agencies, medical institutions, and clinical information system vendors, resulting in widely varying dictionary types; 2. Inefficient system integration; inconsistent terminology across vendors makes data sharing and interoperability difficult; 3. Indiscriminate terminology maintenance by individual medical institutions; no dedicated department for maintenance, and multiple terminology entries exist depending on operational needs, leading to difficulties in later terminology maintenance; 4. A gap exists between existing terminology and clinical application, especially in electronic medical record writing, where terminology often differs from clinical guidelines and standards.

[0004] Therefore, it is necessary to develop a knowledge graph-based method, electronic device, and medium for mapping medical terminology.

[0005] The information disclosed in the background section of this invention is intended only to enhance the understanding of the general background of this invention, and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] This invention proposes a medical terminology mapping method, electronic device, and medium based on knowledge graphs. By combining knowledge graphs with medical terminology mapping, the relevance of terminology attributes is increased, the compatibility of terminology mapping is improved, and the mapping effect of newly added terms is automatically corrected based on knowledge graphs.

[0007] In a first aspect, embodiments of this disclosure provide a medical terminology mapping method based on a knowledge graph, comprising:

[0008] Determine the basic architecture type of the knowledge graph for medical terminology, then assign relevance values ​​to obtain the knowledge graph;

[0009] The format of the mapped term set is standardized and matched with the knowledge graph to obtain the mapping results of the mapped terms.

[0010] After obtaining the mapping results, for new terms in the mapped term set, a semi-automatic mapping is performed based on the knowledge graph to obtain the mapping results corresponding to the new terms.

[0011] Preferably, the basic architecture types of a knowledge graph include term type, term encoding, term name, term relationship, and attribute value;

[0012] The terminology types include disease, drug, examination, test, and surgery.

[0013] Preferably, relevance refers to the degree of matching between term names and attribute values.

[0014] Preferably, the corresponding relevance is assigned to category 1, category 2, and category 3 according to the degree of matching from high to low, wherein the degree of matching for category 1 is greater than that for category 2 and the degree of matching for category 3.

[0015] Preferably, the format standardization process is performed on the mapped term set, and it is matched with the knowledge graph to obtain the mapping results of the mapped terms, including:

[0016] The mapping result is obtained by matching the terms to be mapped with terms in the knowledge graph that are related to class 1.

[0017] If no term with a relevance of class 1 is matched, the mapped term is segmented, and the segmentation results are matched with terms with relevance of class 1 and class 2 in the knowledge graph to obtain the mapping results.

[0018] Preferably, the word segmentation processing methods include forward maximum matching, reverse maximum matching algorithm, bidirectional maximum matching, and setting segmentation flags.

[0019] Preferably, the mapping results are obtained by matching the word segmentation results with terms in the knowledge graph that are classified as category 1 or category 2 in relevance, including:

[0020] For each word segmentation result, it is matched with terms in the knowledge graph that are classified as category 1 or category 2 in relevance. The matching results are weighted and calculated to obtain the relevance score. The matching result with the highest relevance score is taken as the final matching result.

[0021] Preferably, after obtaining the mapping results, for new terms in the mapped term set, semi-automatic mapping is performed based on the knowledge graph to obtain the mapping results corresponding to the new terms, including:

[0022] Match the new term with the term ID in the mapping results and display the matching mapping results;

[0023] If there is no match for the term ID, then match the term name in the mapping results and display the matching mapping results;

[0024] If no matching results are found for the term ID or term name, the new term is segmented, and the segmentation results are matched with terms in the knowledge graph that are related to the term ID (class 1). The matching mapping results are then displayed.

[0025] Secondly, embodiments of this disclosure also provide an electronic device, the electronic device comprising:

[0026] Memory, which stores executable instructions;

[0027] A processor that executes the executable instructions in the memory to implement the knowledge graph-based medical terminology mapping method.

[0028] Thirdly, embodiments of this disclosure also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the knowledge graph-based medical terminology mapping method.

[0029] The methods and apparatus of the present invention have other features and advantages that will be apparent from or will be set forth in detail in the accompanying drawings and following detailed description, which together serve to explain the particular principles of the invention. Attached Figure Description

[0030] The above and other objects, features and advantages of the present invention will become more apparent from the more detailed description of exemplary embodiments of the invention in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same parts.

[0031] Figure 1 A flowchart illustrating the steps of a knowledge graph-based medical terminology mapping method according to an embodiment of the present invention is shown.

[0032] Figure 2 A flowchart of a knowledge graph-based medical terminology mapping according to an embodiment of the present invention is shown.

[0033] Figure 3 A schematic diagram of a semi-automatic mapping of medical terms according to an embodiment of the present invention is shown. Detailed Implementation

[0034] Preferred embodiments of the invention will now be described in more detail. While preferred embodiments of the invention are described below, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0035] To facilitate understanding of the solutions and effects of the embodiments of the present invention, three specific application examples are given below. Those skilled in the art should understand that these examples are merely for the purpose of understanding the present invention, and any specific details therein are not intended to limit the present invention in any way. Example 1

[0036] Figure 1 A flowchart illustrating the steps of a knowledge graph-based medical terminology mapping method according to an embodiment of the present invention is shown.

[0037] like Figure 1 As shown, this knowledge graph-based medical terminology mapping method includes:

[0038] Step 101: Determine the basic architecture type of the knowledge graph for medical terminology, and then assign relevance values ​​to obtain the knowledge graph;

[0039] Step 102: Perform format standardization processing on the mapped term set, match it with the knowledge graph, and obtain the mapping result of the mapped terms;

[0040] Step 103: After obtaining the mapping results, for new terms in the mapped term set, perform semi-automatic mapping based on the knowledge graph to obtain the mapping results corresponding to the new terms.

[0041] In one example, the basic architecture types of a knowledge graph include term type, term encoding, term name, term relationship, and attribute value;

[0042] The terminology types include disease, drug, examination, test, and surgery.

[0043] In one example, relevance is the degree to which a term name matches an attribute value.

[0044] In one example, the relevance is assigned to class 1, class 2, and class 3 according to the degree of matching from high to low, where the degree of matching for class 1 is greater than that for class 2, which is greater than that for class 3.

[0045] In one example, the format specification of the mapped term set is processed and matched with the knowledge graph to obtain the mapping results of the mapped terms, including:

[0046] The mapping result is obtained by matching the terms to be mapped with terms in the knowledge graph that are related to class 1.

[0047] If no term with a relevance of class 1 is matched, the mapped term is segmented, and the segmentation results are matched with terms with relevance of class 1 and class 2 in the knowledge graph to obtain the mapping results.

[0048] In one example, word segmentation methods include forward maximum matching, reverse maximum matching, bidirectional maximum matching, and setting segmentation markers.

[0049] In one example, matching the word segmentation results with terms in the knowledge graph that are classified as category 1 or 2 in terms of relevance yields the following mapping results:

[0050] For each word segmentation result, it is matched with terms in the knowledge graph that are classified as category 1 or category 2 in relevance. The matching results are weighted and calculated to obtain the relevance score. The matching result with the highest relevance score is taken as the final matching result.

[0051] In one example, after obtaining the mapping results, for new terms in the mapped term set, a semi-automatic mapping is performed based on the knowledge graph to obtain the mapping results corresponding to the new terms, including:

[0052] Match the new term with the term ID in the mapping results and display the matching mapping results;

[0053] If there is no match for the term ID, then match the term name in the mapping results and display the matching mapping results;

[0054] If no matching results are found for the term ID or term name, the new term is segmented, and the segmentation results are matched with terms in the knowledge graph that are related to the term ID (class 1). The matching mapping results are then displayed.

[0055] Specifically, current medical terminology is fragmented, with varying structural types, making maintenance difficult and resulting in inconsistent mapping results. Furthermore, changes to the hospital's terminology dictionary often fail to automatically complete the mapping with new terms. A unified knowledge graph would be beneficial for terminology systems across all systems. Current terminology mapping primarily relies on existing terminology sets, ultimately returning a corresponding result, but this requires that the mapping and mapped terms have largely consistent structures. Considering the significant differences between multiple systems (e.g., substantial differences between knowledge base and clinical terminology), traditional methods fail to address issues such as large and small concepts, grouping tests, and combined concepts, and often return empty results when no complete mapping exists. The inherent logical relationships within a knowledge graph can partially solve these problems. In clinical information system applications, changes to terminology, such as additions or modifications, can render existing mappings invalid or prevent the discovery of new mappings, necessitating manual maintenance. Knowledge graph-based mapping, however, uses not only encoding as the mapping basis but also semantic and knowledge relationship methods to supplement it, achieving semi-automatic graph mapping and reducing the manual cost of terminology mapping maintenance.

[0056] Determine the basic architecture type of the knowledge graph, including term type, term encoding, term name, term relationship, attribute value, and relevance. Construct different types of term types, including diseases, drugs, examinations, tests, surgeries, etc. Construct different types of term relationships and attribute values, including standard names, aliases, categories, and unique term relationships.

[0057] The importance of the term relationship values ​​is divided into three categories: Category 1 attributes are important attributes, and a match indicates a complete match; Category 2 attributes are auxiliary attributes, which need to be matched after the terms are broken down and the relevance is mapped based on the matching degree of different parts; Category 3 attributes are relevant attributes, which further supplement and extend the mapping results, as shown in Table 1, to obtain the knowledge graph.

[0058] Table 1

[0059] .

[0060] Figure 2 A flowchart of a knowledge graph-based medical terminology mapping according to an embodiment of the present invention is shown.

[0061] Based on the knowledge graph, the mapping is compared with the mapped term set (term set A). Word segmentation technology is used to break down the terms in term set A, and then compared with the terms and term attributes in the knowledge graph. Feature word matching is used to obtain the most relevant matching item. If no final matching item is found, related or similar items are determined by associating with higher-level categories (diseases, test kits, examination sites). Finally, the mapping results are manually reviewed, and the results are used for knowledge graph optimization. Figure 2As shown.

[0062] The terminology set A is formatted to remove interfering structures and special punctuation marks. Terms are matched against Class 1 attributes of the knowledge graph; if the result is true, the mapped term code and term name are labeled.

[0063] If no direct mapping exists, the term set A is segmented. Segmentation methods include forward maximum matching, backward maximum matching, bidirectional maximum matching, and segmentation marker methods. Considering that medical terms are often short texts, this method uses bidirectional maximum matching. For example, carotid endarterectomy can be segmented into "carotid\endarterial\stripping surgery". The segmentation results are compared with the knowledge graph's Class 1 and Class 2 attributes. Class 1 attributes are compared first to determine the term's main concept, followed by comparison with Class 2 attributes. The results from Class 1 and Class 2 attributes are weighted to obtain a relevance score, which is then ranked. The term with the highest relevance score is mapped. For example, carotid endarterectomy can be directly mapped to Class 1 attributes, with a matching degree of 100. After segmentation, carotid endarterectomy can be divided into two terms: carotid artery and stripping, each with a matching degree of 10, which, after weighting, is 20. Therefore, carotid endarterectomy is directly mapped to Class 1 attributes, as shown in Table 2.

[0064] Table 2

[0065] type Hospital ID Hospital Name Term id Term name disease Yyjb0001 Cerebral infarction JB0001 Cerebral infarction disease Cerebral infarction Cerebral infarction JB0001 Cerebral infarction drug Yyyp0001 Aspirin tablets YP0001 Aspirin tablets drug Aspirin tablets Aspirin tablets YP0001 Aspirin tablets examine Yyjc0001 Head CT scan JC0001 Head CT scan examine Head CT scan Head CT scan JC0001 Head CT scan Operation Yyss0001 Carotid endarterectomy SS0001 Carotid endarterectomy Operation Carotid endarterectomy Carotid endarterectomy SS0001 Carotid endarterectomy .

[0066] If no mapping result is found, then based on different term types, consider hypernyms and supernyms. For example, for diseases, use a hypernym-filling method to perform mapping. The mapping results are manually reviewed, and erroneous mapping results are used for knowledge graph optimization.

[0067] Figure 3 A schematic diagram of a semi-automatic mapping of medical terms according to an embodiment of the present invention is shown.

[0068] like Figure 3 As shown, in the application of term mapping, based on the knowledge graph and mapping results, semi-automatic mapping can be performed on the dynamic term set A, thereby ensuring that matching can be performed automatically after terms are added or modified, thus increasing the fault tolerance rate.

[0069] The medical terminology mapping results are imported into the dictionary management system backend, and the results are output by calling the terms during the clinical information system call process.

[0070] During the system call, matching is performed based on the ID of terminology set A. If a mapping relationship exists between the ID of terminology set A and the corresponding knowledge graph in the dictionary backend, the mapping result is used to display the content. If no direct mapping relationship exists, the name of terminology set A is used for matching. If a mapping relationship exists between the name of terminology set A and the corresponding knowledge graph in the dictionary backend, the mapping result is used to display the content. This method is compatible with adding new standardized terms and can output the corresponding mapping result without manual operation.

[0071] If there is still no mapping relationship, the term name will be converted into a search term, the keyword stem will be extracted, the combination will be performed, and the search results will be returned.

[0072] If the mapping result is still empty, feedback is provided through the feedback page, and the knowledge graph is manually evaluated and optimized. After the knowledge graph is updated, it will take effect immediately.

[0073] This disclosure provides an electronic device comprising: a memory storing executable instructions; and a processor executing the executable instructions in the memory to implement the aforementioned knowledge graph-based medical terminology mapping method.

[0074] An electronic device according to an embodiment of the present disclosure includes a memory and a processor.

[0075] This memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0076] The processor may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In one embodiment of this disclosure, the processor is used to execute computer-readable instructions stored in the memory.

[0077] Those skilled in the art will understand that, in order to solve the technical problem of how to achieve a good user experience, this embodiment may also include well-known structures such as communication buses and interfaces, and these well-known structures should also be included within the protection scope of this disclosure.

[0078] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0079] This disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the knowledge graph-based medical terminology mapping method.

[0080] A computer-readable storage medium according to embodiments of the present disclosure stores non-transitory computer-readable instructions. When these non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the methods described in the foregoing embodiments of the present disclosure are performed.

[0081] The aforementioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or portable hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

[0082] Those skilled in the art should understand that the above description of the embodiments of the present invention is only intended to illustrate the beneficial effects of the embodiments of the present invention, and is not intended to limit the embodiments of the present invention to any of the examples given.

[0083] The various embodiments of the present invention have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A medical terminology mapping method based on knowledge graphs, characterized in that, include: Determine the basic architecture type of the knowledge graph for medical terminology, then assign relevance values ​​to obtain the knowledge graph; The format of the mapped term set is standardized and matched with the knowledge graph to obtain the mapping results of the mapped terms. After obtaining the mapping results, for new terms in the mapped term set, semi-automatic mapping is performed based on the knowledge graph to obtain the mapping results corresponding to the new terms. The basic architecture types of knowledge graphs include term type, term ID, term name, term relationship, and attribute value. The terminology types include disease, drug, examination, test, and surgery; Relevance refers to the degree of matching between term names and attribute values; The relevance is divided into three categories, 1, 2, and 3, based on the degree of matching from high to low. The degree of matching for category 1 is greater than that for category 2, which is greater than that for category 3. After obtaining the mapping results, for new terms in the mapped terminology set, a semi-automatic mapping is performed based on the knowledge graph to obtain the mapping results corresponding to the new terms, including: Match the new term with the term ID in the mapping results and display the matching mapping results; If there is no match for the term ID, then match the term name in the mapping results and display the matching mapping results; If there are no matching results for the term ID and term name, the new term is segmented, and the segmentation results are matched with terms in the knowledge graph that are related to the term ID and term name. The matching mapping results are then displayed. This process involves format standardization of the mapped term set, matching it with the knowledge graph, and obtaining the mapping results for the mapped terms, including: The mapping result is obtained by matching the terms to be mapped with terms in the knowledge graph that are related to class 1. If no term with a relevance of class 1 is matched, the mapped term is segmented, and the segmentation results are matched with terms with relevance of class 1 and class 2 in the knowledge graph to obtain the mapping results. If no mapping result is found, mapping will be performed using a superordinate class supplementation method based on different term types. The mapping results will be manually reviewed, and erroneous mapping results will be used for knowledge graph optimization. Specifically, the mapping results are obtained by matching the word segmentation results with terms in the knowledge graph that are classified as category 1 or category 2 in terms of relevance. For each word segmentation result, it is matched with terms in the knowledge graph that are classified as category 1 or category 2 in relevance. The matching results are weighted and calculated to obtain the relevance score. The matching result with the highest relevance score is taken as the final matching result.

2. The medical terminology mapping method based on knowledge graphs according to claim 1, wherein, Word segmentation methods include forward maximum matching, reverse maximum matching, bidirectional maximum matching, and setting segmentation markers.

3. An electronic device, characterized in that, The electronic device includes: Memory, which stores executable instructions; A processor that executes the executable instructions in the memory to implement the knowledge graph-based medical terminology mapping method of claim 1 or 2.

4. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the knowledge graph-based medical terminology mapping method as described in claim 1 or 2.

Images