CN-122021822-A - Implicit knowledge dominance and structuring processing method, equipment, medium and product of traditional Chinese medicine ancient medical science case

CN122021822ACN 122021822 ACN122021822 ACN 122021822ACN-122021822-A

Abstract

The application relates to the technical field of traditional Chinese medicine information, and discloses a method, equipment, medium and product for explicit and structured treatment of implicit knowledge of traditional Chinese medicine ancient medical records. The method comprises the steps of obtaining standardized medical case corpus according to original medical case data, determining triples according to standardized medical note or comment materials, enabling the triples to be used for mining implicit knowledge, obtaining an explicit knowledge set according to the triples, enabling the explicit knowledge set to be used for representing existing knowledge and implicit knowledge in a traditional Chinese medical case, obtaining a traditional Chinese medical case knowledge graph according to the explicit knowledge set, obtaining a fusion knowledge base according to the traditional Chinese medical case knowledge graph, and achieving medical case knowledge service and visual application according to the fusion knowledge base.

Inventors

LUO XIAOSHU
CHEN XIAO
WANG YONGJIA
WANG GANG

Assignees

上海信投智能科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20251222

Claims (10)

1. The implicit knowledge dominance and structuring processing method for the ancient medical proposal of the traditional Chinese medicine is characterized by comprising the following steps of: obtaining standardized medical case corpus according to the original medical case data; determining a triplet according to the standardized doctor note or comment material, wherein the triplet is used for mining implicit knowledge; Obtaining an explicit knowledge set according to the triples, wherein the explicit knowledge set is used for representing the existing knowledge and implicit knowledge in the traditional Chinese medicine medical science; According to the explicit knowledge set, obtaining a knowledge graph of the ancient medical case of the traditional Chinese medicine; Aiming at the knowledge graph of the traditional Chinese medicine ancient medical case, a fused complete traditional Chinese medicine knowledge base is obtained; and according to the integrated traditional Chinese medicine knowledge base, medical case knowledge service and visual application are realized.
2. The method of claim 1, wherein the determining a triplet from the standardized medical proposal corpus comprises: Establishing a three-level dynamic semantic annotation system, wherein the three-level dynamic semantic annotation system comprises an L0 level, an L1 level and an L2 level, wherein the L0 level is a basic entity layer and is used for identifying and annotating noun entities existing physically, the L1 level is a clinical semantic layer and is used for annotating the entities of the L0 level to judge clinical roles, and the L2 level is a hidden attribute layer and is used for linking an external traditional Chinese medicine ontology library for the extracted L0 level and L1 level entities and acquiring associated hidden attributes; According to a three-level dynamic semantic annotation system, model-expert collaborative annotation is performed on standardized doctor note or comment materials, and the annotation results of the L0 level and the L1 level are determined; according to the labeling results of the L0 level and the L1 level, mounting the L2 level to obtain a complete three-level labeling result; and extracting and determining the triples according to the complete three-level labeling result, wherein the triples comprise real triples and attribute triples.
3. The method of claim 2, wherein the performing model-expert collaborative labeling for standardized doctor note or comment materials, determining labeling results for L0 and L1 stages, comprises: pre-labeling the L0 level and the L1 level according to the fine tuning model; Calculating the confidence entropy value of each pre-labeling entity; and determining labeling results of the L0 level and the L1 level according to the confidence entropy value.
4. The method of claim 3, wherein the confidence entropy value is calculated by: Wherein, the The confidence entropy value is represented as, The average entropy of the entity is represented, and is the average value of the entropy of the token level, the average level of all token uncertainties in the entity is reflected, and the token represents the minimum text unit of the entity, namely the single word or phrase fragment split by the entity in the marked L0 and L1. The minimum confidence of an entity is represented, is the minimum value of the probability of the maximum category in all the token of the entity, and reflects the confidence level of the token which is the least determined by the entity. Represents the balance coefficient, and the value range is Weight duty cycle for adjusting entity average entropy and entity minimum confidence. Representing a single entity to be calculated in L0 and L1.
5. The method of claim 1, wherein deriving the explicit knowledge-set from the triples comprises: Constructing data in the triplets into initial heterograms, wherein the initial heterograms are used for reflecting explicit association of the triplets, and nodes comprise symptoms, prescription and diagnosis and treatment link faults between the symptoms and prescription; Obtaining an enhanced map through a two-channel reasoning mechanism according to the initial heterogram, wherein the enhanced map comprises a complete diagnosis and treatment path; and applying a hierarchical community discovery algorithm to the enhancement map to obtain a dominant knowledge set.
6. The method of claim 5, wherein the obtaining the enhanced atlas from the initial iso-composition via a two-channel inference mechanism comprises: calculating a first cosine similarity of a symptom vector and a potential symptom vector in an initial heterogram and/or calculating a second cosine similarity of the symptom vector and a treatment vector by using a data driving channel to obtain a potential hidden node list; Utilizing a rule channel, executing an inference mechanism based on a traditional Chinese medicine ontology library, and screening the potential hidden node list to obtain a screened node list; taking intersection of nodes screened by the rule channels and nodes generated by the data channels, and determining hidden nodes; instantiating hidden nodes and inserting the hidden nodes into the fault diagnosis and treatment link of the initial heterogram to obtain an enhanced map.
7. The method of claim 5, wherein applying a hierarchical community discovery algorithm to the enhanced graph results in an explicit knowledge-set, comprising: According to the enhancement map, a hierarchical community algorithm is applied to construct a community map with a multi-level structure, wherein the community map is used for displaying the association mode and the topic clustering characteristic of the clinical knowledge of the traditional Chinese medicine; Digging a recessive drug pair from a bottom layer of the multi-level structure, digging a square group from a middle layer of the multi-level structure, and digging a macroscopic diagnosis and treatment mode from a high layer of the structure; and integrating the dug hidden medicine pair, prescription group and macroscopic diagnosis and treatment mode to obtain a dominant knowledge set.
8. An electronic device, the electronic device comprising: Field programmable gate array, and A memory storing configuration data, the field programmable gate array being configured with the configuration data to form a hardware logic circuit to perform the steps of the method of any one of claims 1 to 7.
9. A computer readable medium having configuration data stored thereon, wherein the configuration data, when loaded into a field programmable gate array, causes the internal configuration of the field programmable gate array to form logic circuitry, performing the steps of the method of any of claims 1 to 7.
10. A computer program product comprising hardware description language code or netlist, characterized in that the hardware description language code or netlist is generated with configuration data for configuring a field programmable gate array to perform the steps of the method of any one of claims 1 to 7.

Description

Implicit knowledge dominance and structuring processing method, equipment, medium and product of traditional Chinese medicine ancient medical science case Technical Field The application relates to the technical field of traditional Chinese medicine information, in particular to a method, equipment, medium and product for explicit and structured treatment of implicit knowledge of the ancient medical records of traditional Chinese medicine. Background In the field of traditional Chinese medicine, the ancient Chinese medicine case has higher research value and clinical application value, and in recent years, the specific application work of the ancient Chinese medicine case is approximately subjected to the stages of digitizing the ancient Chinese medicine case, constructing a knowledge base knowledge graph by depending on the digitalized work, and developing intelligent reasoning questions and answers, so that auxiliary support can be provided for the diagnosis and treatment of the traditional Chinese medicine. The inventor finds that at least the following technical problems are that 1, an inference mechanism is lacked, implicit knowledge mining depth is insufficient, implicit etiology, pathogenesis, rules, compatibility rules and other empirical regular knowledge are difficult to extract from an ancient medical case text, 2, language expression is not standard and complex, processing accuracy is low due to the fact that a large number of terms such as foreign words, dialects, ancient words and fuzzy expression exist in the ancient medical case, 3, a unified structural labeling system is lacked, knowledge map construction automation degree is low, fusion capability with traditional Chinese medicine theory is weak, 4, knowledge service intelligence level is limited, existing traditional Chinese medicine intelligent auxiliary models are poor in generalization capability and interpretation, complex clinical problems are faced to intelligently provide simple retrieval results, 5, a system processing flow is split, incremental learning and dynamic updating mechanisms are lacked, and higher maintenance cost reconstruction is needed for updating data. Disclosure of Invention The application aims to provide a method for explicit and structured treatment of implicit knowledge of ancient medical cases of traditional Chinese medicine, which is at least used for assisting in solving the technical problems that empirical rule knowledge is difficult to extract from medical case texts, a unified structured labeling system is lacked, the intelligent level of knowledge service is limited and the system treatment flow is split. To achieve the above object, some embodiments of the present application provide the following aspects: According to the first aspect, some embodiments of the application provide a method for explicit knowledge and structuring of a traditional Chinese medicine ancient medical case, which comprises the steps of obtaining standardized medical case corpus according to original medical case data, determining triples according to standardized medical note or comment materials, wherein the triples are used for mining implicit knowledge, obtaining an explicit knowledge set according to the triples, the explicit knowledge set is used for representing existing knowledge and implicit knowledge in the traditional Chinese medicine medical case, obtaining a traditional Chinese medicine ancient medical case knowledge map according to the explicit knowledge set, obtaining a fused complete traditional Chinese medicine knowledge base according to the traditional Chinese medicine ancient medical case knowledge map, and realizing medical case knowledge service and visual application according to the fused complete traditional Chinese medicine knowledge base. In a second aspect, some embodiments of the application also provide an electronic device comprising one or more processors and a memory storing computer program instructions that, when executed, cause the processors to perform the steps of the method as described above. In a third aspect, some embodiments of the application also provide a computer readable medium having stored thereon computer program instructions executable by a processor to implement a method as described above. In a fourth aspect, some embodiments of the application also provide a computer program product comprising a computer program/instruction which, when executed by a processor, implements the steps of the method as described above. Compared with the related technology, in the scheme provided by the embodiment of the application, aiming at the problems of long processing period, high cost and difficult updating of the medical case knowledge service system constructed by the traditional method, the full-flow optimization method from medical case data acquisition and digital processing, preprocessing and ancient text standardization, medical semantic layering labeling and entity extraction, im