Knowledge graph completion method for dhi report interpretation, dhi report interpretation method, storage medium and equipment

Abstract

The application discloses a knowledge graph completion method for DHI report interpretation, a DHI report interpretation method, a storage medium and equipment, and belongs to the technical field of livestock breeding. In order to solve the problem that the knowledge graph for DHI report interpretation at present is constructed based on known information and cannot effectively interpret potential performance indicators / symptoms and influence factors, thereby leading to incomplete DHI report interpretation, the application firstly determines entities and entity relationships by using known relationships based on a basic text library, and constructs a benchmark DHI field knowledge graph in combination with a crowdsourcing calculation mode; then, the application completes the knowledge graph by predicting entity relationships based on a big prophecy model, determines the edge weight of an incremental knowledge graph by using objectively existing logical relationship representation, and adjusts the edge weight of the basic knowledge graph, so that the application can construct a knowledge graph capable of containing potential performance indicators / symptoms and influence factors.

Description

Technical Field

[0001] This invention belongs to the field of animal husbandry technology, and relates to a knowledge graph completion method for DHI report interpretation, as well as a DHI report interpretation method, storage medium, and device. Background Technology

[0002] DHI (Dairy Herd Improvement) refers to a milk production performance testing system. It mainly measures milk yield and milk composition in dairy cows, collects herd data, and analyzes the data to regularly generate testing guidance reports. The information in these reports guides dairy farm feeding, breeding, and management. Correct interpretation of DHI reports can provide effective guidance for dairy herd feeding and management.

[0003] Currently, DHI reports are primarily interpreted by experienced experts. However, not all farms are equipped with such experts, meaning most farms cannot interpret DHI reports correctly and effectively. This prevents them from fully utilizing the value of DHI reports and from providing scientific and effective guidance for dairy herd management. Furthermore, differences in expertise among specialists lead to variations in DHI report interpretation, hindering the development of a unified standard. This prevents farms from establishing a consistent management system and raises questions about the scientific rigor and effectiveness of dairy herd management practices.

[0004] Gao Meng et al. from Northeast Agricultural University proposed a "DHI report interpretation method based on knowledge graph (application number CN202110969609.X)". This method combines the DHI domain knowledge graph to diagnose problems based on the results of dynamic analysis. Based on the DHI domain knowledge graph, which includes three types of entities and entity relationships: "performance indicators / symptoms", "influencing factors", and "solutions", it realizes the automatic interpretation of DHI reports, enabling any dairy herd management unit to receive scientific and effective guidance. At the same time, it forms a unified standard for DHI report interpretation, enabling different farm management to form a unified system and ensuring the scientific and effective nature of dairy herd management. Summary of the Invention

[0005] This invention aims to address the problem that current knowledge graphs used for DHI report interpretation are built based on known information and cannot effectively interpret potential performance indicators / symptoms and influencing factors, resulting in incomplete interpretation of DHI reports.

[0006] The knowledge graph completion method for DHI report interpretation includes the following steps:

[0007] First, obtain the baseline DHI domain knowledge graph, and denote the weights of the edges in the baseline DHI domain knowledge graph as the basic edge weights. Then, based on the benchmark DHI domain knowledge graph, the knowledge is completed, including the following steps:

[0008] S201. Using all entities in the baseline DHI domain knowledge graph as reference entities, extract entities and entity relationships from all texts in the incremental text library based on the language big model, and predict tail entities through the language big model to obtain predicted triples (h, r, t), where h and t represent the head entity and tail entity in the triple, and r represents the head entity relationship in the triple; complete the baseline DHI domain knowledge graph structure based on the predicted triples to obtain the complete DHI domain knowledge graph structure, denoted as the full DHI domain knowledge graph, and the difference between the full DHI domain knowledge graph and the baseline DHI domain knowledge graph is denoted as the incremental DHI domain knowledge graph.

[0009] In the baseline DHI domain knowledge graph, the neighboring nodes of a given node are denoted as basic connection nodes, and their corresponding edges are basic edges. In the incremental DHI domain knowledge graph, the neighboring nodes are denoted as incremental connection nodes, and their corresponding edges are incremental edges. For a given node, the ratio of its corresponding incremental edges to the total number of edges is calculated. , as the incremental adjustment coefficient; 1- As a basic adjustment coefficient;

[0010] S202. Based on the DHI knowledge graph, which contains two types of entities: "influencing factors" and "performance indicators / symptoms," count the unique triples corresponding to all entities in the incremental text library. ; in all unique triples In the context of a specific entity , will contain entities The triplet is denoted as ,statistics quantity Based on Obtain the entity Corresponding incremental edge weights ;

[0011] S203, Correct the basic edge weights in the construction process of the benchmark DHI domain knowledge graph to And correct the incremental edge weights to This will further complete the baseline DHI domain knowledge graph.

[0012] Furthermore, the construction process of the benchmark DHI domain knowledge graph includes:

[0013] (1) Construct the DHI domain ontology, which includes three types of entities and entity relationships: “performance indicators / symptoms”, “influencing factors”, and “solutions”. The performance indicators / symptoms refer to the indicators or symptoms exhibited by dairy cows to represent their health status.

[0014] (2) The electronic text obtained after digitizing the DHI measurement and application guidance materials is used as the annotation object. The ontology is used as the annotation basis to perform semantic annotation on the electronic text data to form annotation data;

[0015] (3) Using the data in the labeled data as training data, extract entities and entity relations from the text in the basic text library according to the ontology structure of the DHI domain knowledge graph, obtain entity and entity relation data, construct triples of any two types of entities and entity relations, and the DHI domain knowledge graph, which is denoted as the benchmark DHI domain knowledge graph.

[0016] (4) For triples in the DHI knowledge graph containing two types of entities, "influencing factors" and "performance indicators / symptoms", calculate the conditional probability between the two types of entities, denoted as . As the weight of the edge between entities, it is denoted as the base edge weight. .

[0017] Furthermore, the aforementioned Obtained through crowdsourcing calculation.

[0018] Furthermore, based on Obtain the entity Corresponding incremental edge weights The process includes:

[0019] For a certain entity The statistical incremental text library contains Number of documents and calculate entity Corresponding incremental edge weights Where N is the number of documents in the incremental text library.

[0020] Furthermore, the knowledge graph adopts the CompGCN network as its network framework;

[0021] Furthermore, the language big model is selected from the BERT model.

[0022] A method for interpreting DHI reports, including:

[0023] First, obtain the DHI data of the ranch;

[0024] Then, the DHI report is interpreted using the knowledge graph completion method used for DHI report interpretation.

[0025] A computer storage medium storing at least one instruction, which is loaded and executed by a processor to implement the DHI report interpretation method.

[0026] A DHI report interpretation device includes a processor and a memory, the memory storing at least one instruction, which is loaded and executed by the processor to perform the DHI report interpretation method.

[0027] Beneficial effects:

[0028] This invention uses the Big Prophecy model to predict entity relationships and complete the knowledge graph. It also determines the edge weights of the incremental knowledge graph by representing objectively existing logical relationships and adjusts the edge weights of the basic knowledge graph. This invention can construct a knowledge graph that can contain potential performance indicators / symptoms and influencing factors, and can interpret these potential performance indicators / symptoms and influencing factors. Therefore, the interpretation of DHI reports is more comprehensive and effective. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the DHI domain ontology.

[0030] Figure 2 A schematic diagram illustrating the process of completing a knowledge graph.

[0031] Figure 3 This is a schematic diagram of the report interpretation process based on a complete knowledge graph. Detailed Implementation Specific implementation method one:

[0033] This embodiment is a DHI report interpretation method, which is actually a DHI report interpretation method based on knowledge graph completion. It includes two parts: one part is a knowledge graph completion method for DHI report interpretation, and the other part is a DHI report interpretation method using the completed knowledge graph.

[0034] First, the knowledge graph used for interpreting DHI reports is completed, including the following steps:

[0035] S1. Construction of the benchmark DHI domain knowledge graph:

[0036] (1) Construct the DHI domain ontology. The DHI domain ontology includes three types of entities and entity relationships: "performance indicators / symptoms", "influencing factors", and "solutions". Figure 1 As shown; the performance indicators / symptoms refer to the indicators or symptoms exhibited by dairy cows to represent their health status.

[0037] (2) The electronic text obtained by digitizing professional books and documents related to DHI reports and application guidelines is used as the annotation object. The ontology is used as the annotation basis to perform semantic annotation on the electronic text data to form annotation data;

[0038] (3) Using the labeled data as training data, entities and entity relations are extracted from the text in the basic text library according to the ontology structure of the DHI domain knowledge graph. Supervised, semi-supervised and unsupervised methods are used to obtain entity and entity relation data. Triples of any two types of entities and entity relations are constructed, as well as the DHI domain knowledge graph, which is denoted as the baseline DHI domain knowledge graph. The knowledge graph uses the CompGCN network as the network framework. The basic text library includes professional books, documents and DHI reports on dairy cows and pastures, as well as texts from the Internet such as Baidu Encyclopedia. It can be constructed according to the actual situation. It should be noted that the more text in this basic text library, the better the effect of constructing the baseline DHI domain knowledge graph will be in theory, but the workload will increase greatly. This invention can construct a basic text library of appropriate size, but it needs to be as comprehensive as possible to cover all known entities and entity relations.

[0039] (4) For triples in the DHI knowledge graph containing two types of entities, "influencing factors" and "performance indicators / symptoms", calculate the conditional probability between the two types of entities, denoted as . As the weight of the edge between entities, it is denoted as the base edge weight. .

[0040] The data is obtained through crowdsourcing. Specifically, crowdsourcing software provides all potential influencing factors for a performance indicator / symptom to participants (farm production personnel, farm managers, domain experts, etc.). Each participant ranks and scores the degree of influence of these factors. Based on the data, the weight of each influencing factor with respect to a performance indicator / symptom is calculated according to the principle of ranking first and then scoring. That is, the most frequent influencing factor in each position is selected as the influencing factor for that position, and the average score of the participants for that influencing factor in that position is used as the weight between that influencing factor and the performance indicator / symptom. The same method is used to obtain the correlation strength between each performance indicator / symptom and each influencing factor, forming a correlation coefficient matrix, which represents the weight of the edges between the two types of entities: performance indicator / symptom and influencing factor.

[0041] S2, Completion of the benchmark DHI domain knowledge graph:

[0042] Existing professional books, literature, and DHI reports on dairy cows and pastures use labeled data to identify entities and their relationships based on existing knowledge. For example, entities and their relationships are only identified and extracted after it is known which "influencing factors" affect a certain "performance indicator / symptom." Supervised, semi-supervised, and even unsupervised methods for extracting entities and their relationships are also largely based on this "known" information. However, situations exist where "objectively, entities and their relationships actually exist, but we are not currently aware of them." This leads to an incomplete knowledge graph, limiting its interpretability. Therefore, this implementation focuses on supplementing the baseline DHI domain knowledge graph:

[0043] S201. Using all entities in the benchmark DHI domain knowledge graph as reference entities, extract entities and entity relations (h, r, ?) from all texts in the incremental text library based on the BERT model, and predict tail entities through the BERT model to obtain the predicted triples (h, r, t).

[0044] Based on the predicted triples, the baseline DHI domain knowledge graph structure is completed to obtain the complete DHI domain knowledge graph structure, which is denoted as the full DHI domain knowledge graph. The difference between the full DHI domain knowledge graph and the baseline DHI domain knowledge graph is denoted as the incremental DHI domain knowledge graph.

[0045] In the baseline DHI domain knowledge graph, the neighboring nodes of a given node are denoted as basic connection nodes, and their corresponding edges are basic edges. In the incremental DHI domain knowledge graph, the neighboring nodes are denoted as incremental connection nodes (actually, the nodes that have been completed), and their corresponding edges are incremental edges. For a given node, the ratio of its corresponding incremental edges to the total number of edges is calculated. , as the incremental adjustment coefficient; 1- As a basic adjustment coefficient. For example, a node in the baseline DHI domain knowledge graph has four neighboring nodes. After completion, there are a total of six nodes, of which the two newly added nodes are incremental connection nodes. The edges between this node and the incremental connection nodes are incremental edges. For this node, the incremental adjustment coefficient is... The basic adjustment coefficient is .

[0046] S202. Based on the DHI knowledge graph, which contains two types of entities: "influencing factors" and "performance indicators / symptoms," count the unique triples corresponding to all entities in the incremental text library. ; in all unique triples In the context of a specific entity , will contain entities The triplet is denoted as ,statistics quantity During this process, the entity As the head entity Tail-end entity The number of corresponding triples must be counted;

[0047] For a certain entity The statistical incremental text library contains Number of documents and calculate entity Corresponding incremental edge weights Where N is the number of documents in the incremental text library.

[0048] The essence of the process of completing the baseline DHI domain knowledge graph is to supplement which entities that objectively exist but whose relationships are not yet known. It is precisely because the relationships between entities are unknown that the weights of the edges cannot be obtained through step (4). In order to determine that the weights of the edges cannot be obtained through (4), this invention turns to mining the objectively existing conditions. Since the texts that exist in the incremental text library objectively record the relationships between entities, although some relationships are unknown, they objectively exist. Therefore, based on the BERT model, these unknown but actually existing relationships in the incremental text library can be extracted. And a certain entity in the incremental text library The number of corresponding texts can, to a certain extent, characterize the strength of the objective coexistence of entities and their relationships. Therefore, as long as the number of texts in the incremental text library is sufficient, the relationship can be represented. This invention uses objectively existing phenomena to calculate incremental edge weights. In calculating edge weights, considering that the higher the probability of a certain objective potential relationship, the higher its probability of occurrence, or the more frequently it is potentially "represented," this invention uses the document frequency (the proportion of entity documents) corresponding to a certain entity relationship as its objective potential probability. Simultaneously, considering that for the potential relationships extracted by the large language model, this potential relationship is essentially an "estimation result," that is, an entity is estimated to have entity relationships with other entities. The more relationships an entity has with other entities, the weaker its representational ability, and therefore the lower its relationship confidence. For example, an entity... With ten entities - The existence of an entity relationship indicates that there may be ten influencing factors that could lead to the appearance of a certain symptom or an abnormality in a certain indicator; if the entity... Only with 3 entities - The existence of an entity relationship indicates that only three influencing factors will cause an abnormality in a certain symptom or indicator. Therefore, for the estimated potential relationship, compared to "one symptom / indicator is related to ten influencing factors," "one symptom / indicator is related to three influencing factors" indicates a stronger representativeness of the symptom / indicator and a higher confidence level. Therefore, this invention adopts... quantity The reciprocal represents the confidence level and is used in the calculation of edge weights.

[0049] S203, Correct the basic edge weights in the construction process of the benchmark DHI domain knowledge graph to And correct the incremental edge weights to ;

[0050] This completes the baseline DHI domain knowledge graph completion. The knowledge graph completion process is as follows: Figure 2 As shown. Then, the DHI report is interpreted using the completed DHI domain knowledge graph. The interpretation process follows the existing knowledge graph-based DHI report interpretation methods, as follows: Figure 3 As shown, in some embodiments, the interpretation process includes:

[0051] S301 and DHI indicator data acquisition:

[0052] To obtain DHI data from dairy farms, two types of DHI data are needed: monthly data and historical data. Monthly data is obtained directly from the DHI report file produced by the DHI Testing Center based on the China Dairy Cattle Production Performance Measurement and Analysis System (CNDHI). The data is automatically extracted by software. This method is simple and efficient, and can provide basic measurement data and related statistical indicators, such as average calving interval, lactation days, milk fat percentage, protein percentage, fat-to-protein ratio, peak milk production, peak day, sustainability, and urea nitrogen. Historical data is obtained by software traversing the relevant indicator data in historical DHI report files, extracting the relevant indicator data according to the predetermined fields, and storing it in the database.

[0053] S302. Statistical Analysis of Indicator Data:

[0054] The analysis of indicator data includes both static and dynamic analysis.

[0055] Static analysis involves identifying abnormal indicators based on the monthly data for each indicator, according to the normal range standards for each indicator, and then forming a corresponding factual description.

[0056] Dynamic analysis combines the monthly and historical data of each indicator to analyze the recent changing patterns of each indicator and form a corresponding factual description.

[0057] S303, Problem Diagnosis:

[0058] By combining the knowledge graph of the DHI domain with the results of dynamic analysis, problem diagnosis is performed, which includes two aspects: problem identification and guidance and suggestions.

[0059] Problem localization is based on the DHI domain knowledge graph, treating the dynamically analyzed factual descriptions as "performance indicators / symptoms" entities, and calculating the probability that the factual description is caused by a certain influencing factor, denoted as . :

[0060]

[0061] in, It is the conditional probability between performance indicators / symptoms and influencing factors, that is, the weight of the edges between entities (modified basic edge weights and incremental edge weights). The prior probability of performance indicators / symptoms. The initial values ​​were calculated based on historical DHI reports and pasture record data. Specifically, the prior probability of a performance indicator is the proportion of historical DHI reports showing an abnormality in that indicator out of the total number of reports; the prior probability of a symptom is the proportion of pasture record data showing that symptom out of the total number of pasture cows. Furthermore, It can be updated monthly;

[0062] The guidance and recommendations are based on the influencing factors identified through positioning. By analyzing the relationship between the two types of entities, "influencing factors" and "solutions," from the DHI domain knowledge graph, the corresponding solutions for each influencing factor are found and then fed back to the user. Specific Implementation Method Two:

[0064] The method for interpreting DHI reports using a completed DHI domain knowledge graph is coded to obtain a corresponding interpretation program; then it is stored in a computer storage medium of this embodiment, wherein the storage medium stores at least one instruction, which is loaded and executed by a processor to implement the DHI report interpretation method.

[0065] It should be understood that the instructions include computer program products, software, or computerized methods corresponding to any method described in this invention; the instructions can be used to program computer systems or other electronic devices. Computer storage media may include readable media on which instructions are stored, and may include, but are not limited to, magnetic storage media, optical storage media; magneto-optical storage media include read-only memory (ROM), random access memory (RAM), erasable programmable memory (e.g., EPROM and EEPROM), and flash memory layers, or other types of media suitable for storing electronic instructions. Specific implementation method three:

[0067] The method for interpreting DHI reports using a completed DHI domain knowledge graph is coded to obtain a corresponding interpretation program; then it is stored in a memory and run with a DHI report interpretation device as described in this embodiment. The device includes a processor and a memory. It should be understood that this includes any device including a processor and a memory as described in this invention. The device may also include other units and modules that perform display, interaction, processing, control and other functions through signals or instructions.

[0068] The memory stores at least one instruction, which is loaded and executed by the processor to implement the DHI report interpretation method.

[0069] Those skilled in the art will understand that at least one stored instruction constitutes a computer program product corresponding to a method or system. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0070] This application is described with reference to flowchart illustrations and / or block diagrams of methods, systems, and computer program products according to embodiments of this application, and can also be used with corresponding devices. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0071] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0072] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0073] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0074] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0075] This invention may have other embodiments. Without departing from the spirit and essence of this invention, those skilled in the art can make various corresponding changes and modifications according to this invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. A knowledge graph completion method for DHI report interpretation, characterized in that, The method comprises the following steps: First, a benchmark DHI field knowledge graph is obtained, and the weight of the edge in the benchmark DHI field knowledge graph is denoted as the basic edge weight ω m1 Then, the benchmark DHI field knowledge graph is completed based on the benchmark DHI field knowledge graph, including the following steps: S201, all entities in the benchmark DHI field knowledge graph are taken as reference entities, entities and entity relationships in all texts in the incremental text library are extracted based on a language large model, and tail entities are predicted through the language large model to obtain predicted triples (h, r, t), h and t are used to represent head entities and tail entities in the triples, and r is used to represent the relationship between the head entities in the triples; the benchmark DHI field knowledge graph structure is completed based on the predicted triples to obtain a complete DHI field knowledge graph structure, denoted as a full DHI field knowledge graph, and the difference between the full DHI field knowledge graph and the benchmark DHI field knowledge graph is denoted as an incremental DHI field knowledge graph; The neighbor nodes of a certain node in the benchmark DHI field knowledge graph are denoted as basic connection nodes, and the corresponding edges are basic edges. The neighbor nodes in the incremental DHI field knowledge graph are denoted as incremental connection nodes, and the corresponding edges are incremental edges. For a certain node, the proportion k of the number of incremental edges corresponding to the node to the total number of edges is calculated m , as an incremental adjustment coefficient; and 1-k m is taken as a basic adjustment coefficient; S202、Based on the DHI knowledge graph containing "influencing factors" and "performance indicators / symptoms" two kinds of entities, the number of all entity corresponding non-repeated triples (h, r, t) in the incremental text library is counted; in all non-repeated triples (h, r, t), for a certain entity v m , the triple containing entity v m is recorded as The number n of m is counted, and then based on n m , the corresponding incremental edge weight ω m of entity v m2 is obtained; S203, the basic edge weight in the construction process of the reference DHI field knowledge graph is corrected to (1-k m )ω m1 , and the incremental edge weight is corrected to k m ω m2 ; and then the reference DHI field knowledge graph completion is completed.

2. The knowledge graph completion method for DHI report interpretation according to claim 1, characterized in that, The construction process of the benchmark DHI field knowledge graph comprises: (1) constructing a DHI field ontology, the DHI field ontology comprising three types of entities and entity relationships of "performance indicators / symptoms", "influencing factors" and "solutions"; the performance indicators / symptoms refer to indicators or symptoms exhibited by a dairy cow to present a healthy state; (2) taking electronic texts obtained after electronicizing DHI measurement and application guidance related materials as annotation objects, and annotating the electronic text data based on the ontology to form annotated data; (3) using data in the annotated data as training data, extracting entities and entity relationships from texts in the basic text library according to the ontology structure of the DHI field knowledge graph to obtain entity and entity relationship data, constructing triples of any two types of entities and entity relationships, and constructing a DHI field knowledge graph, denoted as a benchmark DHI field knowledge graph; (4) For the triples containing both "influencing factors" and "performance indicators / symptoms" in the DHI knowledge graph, the conditional probability between the two types of entities is calculated, denoted as P(fac|sym), as the weight of the edge between the entities, denoted as the basic edge weight ω m1 .

3. The knowledge graph completion method for DHI report interpretation according to claim 2, characterized in that, The P(fac|sym) is obtained by a crowdsourcing calculation method.

4. The knowledge graph completion method for DHI report interpretation according to any one of claims 1 to 3, characterized in that, Based on n m Obtaining entity v m Corresponding incremental edge weight ω m2 The process includes: For an entity v m , the number of documents containing in the incremental text corpus is counted and the corresponding incremental edge weight of entity v m is calculated where N is the number of documents in the incremental text corpus.

5. The knowledge graph completion method for DHI report interpretation according to claim 4, characterized in that, The knowledge graph adopts a CompGCN network as a network framework.

6. The knowledge graph completion method for DHI report interpretation according to claim 5, characterized in that, The language large model selects a BERT model.

7. A method of DHI report interpretation, characterized by, The method comprises the following steps: First, DHI data of a pasture is obtained; Then, the completed knowledge graph is used for DHI report interpretation.

8. A computer storage medium, characterized in that, The storage medium stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the DHI report interpretation method of claim 7.

9. A DHI report interpretation device, characterized by, The device comprises a processor and a memory, and the memory stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the DHI report interpretation method of claim 7.

Images