Knowledge graph ontology structure construction method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a knowledge graph ontology structure construction method and device, equipment and a storage medium, and the method comprises the steps: extracting field characteristic data from a preset business field, adjusting an interaction model, obtaining an adjusted interaction model, extracting a core entity in input data through the adjusted interaction model, and obtaining a knowledge graph ontology structure. Generating triple data; the core entities are clustered, a domain concept system is generated, then an ontology structure of the knowledge graph is constructed, and the ontology structure is detected and corrected through the adjusted interaction model; and generating a corresponding database schema according to the corrected ontology structure, performing comparison verification on the triple data and the database schema, updating the database schema according to a comparison verification result, and obtaining a data loading report based on the updating process, thereby completing the construction of the knowledge graph, remarkably reducing the dependence on field experts and manual annotation, and improving the construction efficiency. And the cross-domain adaptive capacity and the knowledge updating efficiency are improved.

Description

Technical Field

[0001] This disclosure relates to the field of knowledge graph technology, and in particular to a method, apparatus, device and storage medium for constructing a knowledge graph ontology structure. Background Technology

[0002] Knowledge graphs, as a structured semantic knowledge base, have been widely applied in fields such as intelligent question answering, information retrieval, and data analysis. Existing knowledge graph construction technologies mainly fall into two categories: manual construction and semi-automatic construction. With the development of deep learning technology, neural network-based extraction models have made some progress in entity recognition and relation extraction tasks, but limitations remain in dynamic ontology construction and automated graph updates. Knowledge updates rely on manual intervention and require repeating the entire process. Furthermore, the independent nature of each process module and the lack of effective interaction lead to low update efficiency and a high error rate. Summary of the Invention

[0003] To solve the above-mentioned technical problems, or at least partially solve them, this disclosure provides a method, apparatus, device, and storage medium for constructing a knowledge graph ontology structure.

[0004] This disclosure provides a method for constructing a knowledge graph ontology structure, the method comprising: Extract domain characteristic data from a preset business domain, adjust the interaction model based on the domain characteristic data to obtain an adjusted interaction model, use the adjusted interaction model to extract the core entities and corresponding relationships between entities from the input data, and generate triplet data based on the core entities and corresponding relationships between entities. The core entities are clustered to generate a domain concept system, and a mapping relationship between each core entity and the domain concept system is established. The ontology structure of the knowledge graph is constructed based on the domain concept system and the mapping relationship. The ontology structure is then detected and corrected using an adjusted interaction model. A corresponding database schema is generated based on the modified ontology structure. The database schema is compared and verified based on the triple data. The database schema is updated based on the comparison and verification results. A data loading report is obtained based on the update process. The data loading report is parsed, and the task execution sequence is determined based on the parsing results and the modified ontology structure. The execution process of the task execution sequence is then detected to complete the construction of the knowledge graph.

[0005] The method provided in this disclosure includes extracting domain characteristic data from a preset business domain, dynamically adjusting an interaction model based on the domain characteristic data to obtain an adjusted interaction model, using the adjusted interaction model to extract core entities and corresponding inter-entity relationships from input data, and generating triplet data based on the core entities and corresponding inter-entity relationships. Extract corresponding domain characteristic data from the preset business domain, dynamically adjust the prompt word template and inference parameters of the interaction model based on the domain characteristic data, and then optimize and adjust the interaction model to obtain the adjusted interaction model. The input data is preprocessed to generate a standardized text dataset. The standardized text dataset is then combined with a preset business domain and an adjusted interaction model to perform entity recognition and obtain the core entities. The semantic relationships between the core entities are analyzed using the adjusted interaction model. Based on the semantic relationships, the entity relationships between each core entity and other core entities are extracted, and triplet data are generated based on the entity relationships and core entities.

[0006] The method provided in this disclosure includes clustering the core entities to generate a domain concept system, establishing a mapping relationship between each core entity and the domain concept system, constructing an ontology structure of a knowledge graph based on the domain concept system and the mapping relationship, and detecting and correcting the ontology structure using an adjusted interaction model. Cluster analysis is performed on the core entities, and a domain concept system is generated based on the cluster analysis results. A mapping relationship between the domain concept system and the core entities is established. The ontology structure of knowledge graphs is dynamically generated based on mapping relationships and domain concept systems; The logical reasoning database is used to verify whether there are any conflicts in the ontology structure. If there are, a correction scheme is generated using the adjusted interaction model. The conflict is then corrected using the correction scheme to obtain the corrected ontology structure.

[0007] The method provided in this disclosure includes generating a corresponding database schema based on the modified ontology structure, comparing and verifying the triplet data with the database schema, updating the database schema based on the comparison and verification results, and obtaining a data loading report based on the update process. The target database is determined based on domain characteristic data, and a database schema is generated based on the target database and the modified ontology structure. A graph framework is created based on the database schema. The triplet data is imported into the graph framework to obtain the import results. The import results are compared and verified. The database schema is updated based on the comparison and verification results. At the same time, the comparison and verification results and the import results are summarized, and a data loading report is generated using the update process.

[0008] The method provided in this disclosure, wherein parsing the data loading report, determining the task execution sequence based on the parsing results and the modified ontology structure, and detecting the execution process of the task execution sequence to complete the construction of the knowledge graph, includes: Parse the data loading report to obtain the data volume, task priority, and dependencies. Allocate computing resources based on the data volume, task priority, and dependencies, and generate a task execution sequence by modifying the ontology structure. The execution status and progress data of each task in the task execution sequence are collected in real time. The execution status and progress data are weighted by task, and the overall progress of knowledge graph construction is determined based on the task weighting results. Anomaly detection is performed based on the execution status and progress data, and a retry mechanism and anomaly prompt are triggered based on the anomaly detection results. Complete the construction of the knowledge graph.

[0009] This disclosure also provides a knowledge graph ontology structure construction apparatus, the apparatus comprising: The interaction module is used to extract domain characteristic data from a preset business domain, dynamically adjust the interaction model based on the domain characteristic data to obtain an adjusted interaction model, use the adjusted interaction model to extract the core entities and corresponding relationships between entities from the input data, and generate triplet data based on the core entities and corresponding relationships between entities. The construction module is used to cluster the core entities, generate a domain concept system, establish a mapping relationship between each core entity and the domain concept system, construct the ontology structure of the knowledge graph based on the domain concept system and the mapping relationship, and use the adjusted interaction model to detect and correct the ontology structure. The management module is used to generate a corresponding database schema based on the corrected ontology structure, compare and verify the triple data with the database schema, update the database schema based on the comparison and verification results, and obtain a data loading report based on the update process. The detection module is used to parse the data loading report, determine the task execution sequence based on the parsing results and the corrected ontology structure, and detect the execution process of the task execution sequence.

[0010] In some possible implementations, the interaction module is specifically used for: The optimization unit is used to extract corresponding domain characteristic data from a preset business domain, dynamically adjust the prompt word template and inference parameters of the interaction model according to the domain characteristic data, and then optimize and adjust the interaction model to obtain the adjusted interaction model. The recognition unit is used to preprocess the input data, generate a standardized text dataset, and perform entity recognition on the standardized text dataset in combination with a preset business domain and an adjusted interaction model to obtain the core entities. The extraction unit is used to analyze the semantic associations between the core entities using the adjusted interaction model, extract the entity relationships between each core entity and other core entities based on the semantic associations, and generate triplet data based on the entity relationships and core entities.

[0011] In some possible implementations, the building module is specifically used for: Clustering units are used to perform cluster analysis on the core entities, generate a domain concept system based on the cluster analysis results, and establish a mapping relationship between the domain concept system and the core entities. The ontology generation unit is used to dynamically generate the ontology structure of knowledge graphs based on mapping relationships and domain concept systems. The verification unit is used to verify whether there is a conflict in the ontology structure using a logical reasoning database. If there is a conflict, a correction scheme is generated using the adjusted interaction model, and the conflict is corrected using the correction scheme to obtain a corrected ontology structure.

[0012] In some possible implementations, the management module is specifically used for: The adaptation unit is used to determine the target database based on domain characteristic data; The schema generation unit is used to generate a database schema based on the target database and the modified ontology structure. The report generation unit is used to create a graph framework based on the database schema, import the triplet data into the graph framework, obtain the import result, perform a comparison and verification on the import result, update the database schema based on the comparison and verification result, summarize the comparison and verification result and the import result, and generate a data loading report using the update process.

[0013] In some possible implementations, the detection module includes: The sequence generation unit is used to parse the data loading report, obtain the data volume, task priority and dependency relationship, allocate computing resources based on the data volume, task priority and dependency relationship, and generate a task execution sequence in combination with the modified ontology structure; The progress determination unit is used to collect the execution status and progress data of each task in the task execution sequence in real time, perform task weighting on the execution status and progress data, and determine the overall progress of knowledge graph construction based on the task weighting results. An anomaly detection unit is used to perform anomaly detection based on the execution status and progress data, and trigger a retry mechanism and anomaly prompt based on the anomaly detection result. Complete the construction of the knowledge graph.

[0014] This disclosure also provides an electronic device, the electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the knowledge graph ontology structure construction method provided in this disclosure.

[0015] This disclosure also provides a computer-readable storage medium storing a computer program for executing the knowledge graph ontology structure construction method provided in this disclosure.

[0016] The technical solution provided in this disclosure has the following advantages compared with the prior art: The knowledge graph ontology structure construction method provided in this disclosure adjusts the interaction model through domain data, extracts entities and generates triples, generates a domain concept system through clustering to construct the ontology structure, and achieves self-correction of the ontology. Finally, a database schema is generated based on the ontology structure, and after comparison and verification, the task sequence is driven to complete the graph construction. This constructs an automated knowledge graph construction system that can drive itself and optimize itself. From extracting core entities to generating triples, to building the ontology structure, and finally to generating the database schema and verifying it, the whole process is coherent and automated. Through unified scheduling of task execution sequences, an end-to-end solution is formed, which significantly reduces the dependence on domain experts and manual annotation. The adaptive ontology generation and correction mechanism improves cross-domain adaptability and knowledge update efficiency, and ensures graph quality and construction efficiency. Attached Figure Description

[0017] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0018] Figure 1 A flowchart illustrating the knowledge graph ontology structure construction method provided in this embodiment of the disclosure; Figure 2 This disclosure provides an automated knowledge graph construction process. Figure 3 This is a schematic diagram of the knowledge graph ontology structure construction device provided in the embodiments of this disclosure; Figure 4 This disclosure provides an automated knowledge graph construction architecture. Figure 5 This is an automatic spectrum timing flowchart provided in the embodiments of this disclosure; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0019] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0020] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0021] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0022] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0023] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0024] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0025] To address the aforementioned issues, this disclosure provides a method for constructing a knowledge graph ontology structure, which will be described below with reference to specific embodiments.

[0026] Figure 1This is a flowchart illustrating a method for constructing a knowledge graph ontology structure according to an embodiment of the present disclosure. The method can be executed by a knowledge graph ontology structure construction device, which can be implemented in software and / or hardware and is generally integrated into an electronic device.

[0027] Example 1: This embodiment of the present disclosure provides a method for constructing a knowledge graph ontology structure, such as... Figure 1 and Figure 2 As shown, it includes: Extract domain characteristic data from a preset business domain, adjust the interaction model based on the domain characteristic data to obtain an adjusted interaction model, use the adjusted interaction model to extract the core entities and corresponding relationships between entities from the input data, and generate triplet data based on the core entities and corresponding relationships between entities. The core entities are clustered to generate a domain concept system, and a mapping relationship between each core entity and the domain concept system is established. The ontology structure of the knowledge graph is constructed based on the domain concept system and the mapping relationship. The ontology structure is then detected and corrected using an adjusted interaction model. A corresponding database schema is generated based on the modified ontology structure. The database schema is compared and verified based on the triple data. The database schema is updated based on the comparison and verification results. A data loading report is obtained based on the update process. The data loading report is parsed, and the task execution sequence is determined based on the parsing results and the modified ontology structure to complete the construction of the knowledge graph.

[0028] In this embodiment, the preset business domain refers to the specific, pre-defined professional scope or industry category targeted by the knowledge graph construction. For example, semiconductor manufacturing processes.

[0029] In this embodiment, domain-specific data are core data samples extracted from a preset business domain that represent the unique knowledge structure of that domain. These are used to teach the model to understand domain-specific concepts and relationships. For example, when cardiovascular disease treatment is the preset business domain, the corresponding domain-specific data includes key descriptive texts about diseases, symptoms, and drugs from medical textbook chapters, diagnostic guidelines, and authoritative journal articles.

[0030] In this embodiment, the interaction model refers to the unadjusted large language model, whose input is prompt words and text, and whose output is the initially identified entities or relationships. The adjusted interaction model is a model optimized with domain data, whose input is domain-specific prompt words and text, and whose output is more accurate entities or relationships that conform to domain knowledge. For example, before adjustment, the model might identify an apple as a fruit or a company; after adjustment, in the semiconductor field, it can accurately identify the Apple M3 chip as a core entity.

[0031] In this embodiment, the core entity is an identified independent object that is of key significance in the business domain; the relationship between entities is an association or connection with specific semantics between two or more core entities, identified by the adjusted interaction model. For example, core entities include technology companies, 5G baseband chips, and research and development; the relationship between entities may be an action relationship between technology companies and 5G baseband chips, and the specific type of this relationship is research and development.

[0032] In this embodiment, a triplet is a data unit that expresses a complete fact, consisting of two core entities and the entity relationship between them. The format is typically (subject, relation, object). For example, combining the entities and relations above generates a triplet: (nifedipine, treatment, hypertension). This is the most basic knowledge unit in a knowledge graph.

[0033] In this embodiment, clustering analysis is an unsupervised machine learning method used to automatically group a large number of core entities based on the similarity of their attributes and contextual relationships. For example, when the preset business domain is the semiconductor field, the algorithm clusters lithography machines, etching machines, and ion implanters into the semiconductor manufacturing equipment group; and clusters silicon wafers, photoresist, and specialty gases into the semiconductor materials group.

[0034] In this embodiment, the domain concept system is a hierarchical set of categories abstracted from core entities through cluster analysis, defining the concept classification within the domain. For example, when the preset business domain is the semiconductor domain, the automatically generated system includes core concepts such as manufacturing equipment, chip design software, materials, and process nodes. Among them, the sub-concepts of manufacturing equipment include lithography equipment, etching equipment, etc.

[0035] In this embodiment, the mapping relationship is an instance-category association established between the domain concept system and the core entities. For example, when the preset business domain is the semiconductor domain, a mapping is established between the core entity ASML EUV lithography machine and the domain concept manufacturing equipment; and a mapping is established between the entity FinFET and the domain concept chip architecture.

[0036] In this embodiment, the ontology structure of the knowledge graph is a framework that formally defines concept categories, attributes, and relationships between categories; it is an abstract data model of the knowledge graph. For example, if the preset business domain is the semiconductor field, the ontology structure will explicitly define that manufacturing equipment has attributes such as manufacturer and precision; and that there is a support relationship between chip design software and chip architecture.

[0037] In this embodiment, the revised ontology structure is the final version of the ontology structure that is logically consistent after adjustments based on the revision scheme. For example, when the preset business domain is the semiconductor domain, in the revised ontology structure, computational lithography software is clearly classified as a software tool, and all logical contradictions are eliminated by associating it with the lithography equipment through control attributes.

[0038] In this embodiment, the database schema is a specific physical storage blueprint automatically generated based on the target database type and modified ontology structure. It defines data table / node types, attribute field / edge types, and their indexing strategies.

[0039] In this embodiment, the comparison verification is a process of comparing and analyzing the triple data to be stored with the current structure and content of the database schema during the knowledge graph construction process, and updating the database with new triple data identified in the comparison.

[0040] In this embodiment, the data loading report is a summary document that integrates the import results and the comparison and verification results. It comprehensively evaluates the quality, performance, and readiness status of the data loading. For example, when the preset business field is the semiconductor field, the report summarizes: a total of 98,500 data entries were imported, with 15 failures; the average response time for simulated queries was <200ms; and a supplier relationship missing rate of 5% was found, which is recommended for investigation.

[0041] In this embodiment, the execution process refers to the specific running status of each independent task at a certain moment in the task execution sequence. This includes execution status and progress data.

[0042] In this embodiment, the task execution sequence is a dynamically generated, ordered list of tasks based on task dependencies, priorities, and resource allocation, used to systematically guide the construction process. For example, the generated sequence might be: data cleaning → entity identification → relation extraction → quality assessment → index building.

[0043] In this embodiment, task parameters are obtained by parsing the data loading report, a task sequence is generated based on the allocation of computing resources and the correction of the ontology structure, the task status is monitored in real time and the overall progress is calculated by weighting, and a retry mechanism is triggered through anomaly detection to finally complete the map construction.

[0044] The working principle and beneficial effects of the above embodiments are as follows: Interaction models are adjusted using domain data, entities are extracted and triples are generated, a domain concept system is generated through clustering to construct the ontology structure, and self-correction of the ontology is achieved. Finally, a database schema is generated based on the ontology structure, and after comparison and verification, the task sequence is driven to complete the graph construction. This constructs a self-driven, self-optimizing automated knowledge graph construction system. From extracting core entities to generating triples, then to constructing the ontology structure, and finally to generating the database schema and performing verification, the entire process is coherent and automated. Through unified scheduling of task execution sequences, an end-to-end solution is formed, significantly reducing the dependence on domain experts and manual annotation. The adaptive ontology generation and correction mechanism improves cross-domain adaptability and knowledge update efficiency, ensuring graph quality and construction efficiency.

[0045] Example 2: The method provided in this embodiment of the present disclosure includes extracting domain characteristic data from a preset business domain, dynamically adjusting the interaction model based on the domain characteristic data to obtain an adjusted interaction model, using the adjusted interaction model to extract core entities and corresponding inter-entity relationships from the input data, and generating triplet data based on the core entities and corresponding inter-entity relationships, including: Extract corresponding domain characteristic data from the preset business domain, dynamically adjust the prompt word template and inference parameters of the interaction model based on the domain characteristic data, and then optimize and adjust the interaction model to obtain the adjusted interaction model. The input data is preprocessed to generate a standardized text dataset. The standardized text dataset is then combined with a preset business domain and an adjusted interaction model to perform entity recognition and obtain the core entities. The semantic relationships between the core entities are analyzed using the adjusted interaction model. Based on the semantic relationships, the entity relationships between each core entity and other core entities are extracted, and triplet data are generated based on the entity relationships and core entities.

[0046] In this embodiment, the prompt word template and inference parameters form a structured instruction framework that guides the large language model to perform a specific task. The inference parameters are configurations that control the model's generative behavior, such as creativity and determinism. For example, a prompt word template for entity recognition might be "Find all medical symptoms from the following text," where the entities are text. The inference parameters are set to low randomness to ensure stable output.

[0047] In this embodiment, the optimization process utilizes domain-specific data to automatically modify the content of the prompt template and the values ​​of the inference parameters, making the general interaction model more accurately adapt to the preset business domain. For example, using medical data, the prompt "Please find all entities from the following text" in the template is optimized to "Please find all disease, symptom, and drug entities," and the parameters are adjusted to be more biased towards logical reasoning.

[0048] In this embodiment, the input data is multi-source data, including structured data and unstructured data, and the input data is formatted uniformly through a standardized interface.

[0049] In this embodiment, preprocessing involves cleaning (duplicate removal, error correction), word segmentation, and format conversion of the input data to generate a text sequence that can be processed by a large model. The standardized text dataset is a collection of clean and uniformly formatted plain text data obtained after preprocessing. For example, if the original input is a webpage from a medical forum, after preprocessing, advertisements and navigation bars are removed, and pure case discussion text is extracted to form a standardized text dataset.

[0050] In this embodiment, entity recognition is a technique for locating and classifying named elements from standardized text. For example, from a medical record text describing a patient's complaint of hypertension and the use of nifedipine for control, entity recognition identifies the two core entities: hypertension and nifedipine.

[0051] In this embodiment, semantic association refers to the logical connection in meaning between core entities. Entity relationships are association types that are explicitly defined and categorized based on semantic association. For example, there is a semantic association of treatment between the core entities hypertension and nifedipine, from which the explicit entity relationship can be extracted as nifedipine is used to treat hypertension.

[0052] The working principle and beneficial effects of the above embodiments are as follows: Based on domain characteristic data, the prompt word template and inference parameters of the interaction model are dynamically adjusted to generate an optimized model. Entity recognition is performed on standardized text data to obtain core entities. By analyzing the semantic relationships between entities through the model, entity relationships are extracted and triple data is generated, realizing domain-adaptive entity and relationship extraction. This significantly improves the processing accuracy of professional terms and complex semantics. By dynamically optimizing model parameters, the dependence on manually labeled data is reduced, and cross-domain generalization ability is enhanced.

[0053] Example 3: The method provided in this embodiment of the present disclosure includes clustering the core entities to generate a domain concept system, establishing a mapping relationship between each core entity and the domain concept system, constructing an ontology structure of a knowledge graph based on the domain concept system and the mapping relationship, and using an adjusted interaction model to detect and correct the ontology structure, including: Cluster analysis is performed on the core entities, and a domain concept system is generated based on the cluster analysis results. A mapping relationship between the domain concept system and the core entities is established. The ontology structure of knowledge graphs is dynamically generated based on mapping relationships and domain concept systems; The logical reasoning database is used to verify whether there are any conflicts in the ontology structure. If there are, a correction scheme is generated using the adjusted interaction model. The conflict is then corrected using the correction scheme to obtain the corrected ontology structure.

[0054] In this embodiment, the logical reasoning database is a knowledge base or engine with logical reasoning capabilities, which can automatically discover contradictions based on rules. For example, when the preset business domain is the semiconductor field, the database has pre-defined rules that physical devices and software tools are mutually exclusive categories.

[0055] In this embodiment, the verification process involves using a logical reasoning database to automatically check the generated ontology structure to identify logical conflicts. For example, if the preset business domain is the semiconductor domain, the verification process may find that the ontology declares that physical devices and software tools are mutually exclusive, but simultaneously maps physical computational lithography software to both classes, triggering a logical conflict.

[0056] In this embodiment, the correction scheme is a modification suggestion generated by the adjusted interaction model based on logical conflicts. For example, when the preset business domain is the semiconductor domain, the correction scheme generated by the model for the above-mentioned conflicts is: to change the category of computational lithography software to software tool, and to add control attributes to it to associate it with lithography equipment.

[0057] The working principle and beneficial effects of the above embodiments are as follows: cluster analysis is performed on core entities to generate a domain concept system, establishing a mapping relationship between them and entities and dynamically constructing an ontology structure; ontology conflicts are verified through a logical reasoning database, and a correction scheme is generated and executed using an adjusted interaction model, ultimately forming a corrected ontology structure. This achieves automated construction and dynamic optimization of the ontology structure, significantly reducing reliance on manual domain analysis. Through logical conflict detection and intelligent correction mechanisms, the consistency of ontology logic is ensured, improving the quality and maintainability of the knowledge graph.

[0058] Example 4: The method provided in this embodiment of the present disclosure, wherein the step of generating a corresponding database schema based on the modified ontology structure, comparing and verifying the triplet data with the database schema, updating the database schema based on the comparison and verification results, and obtaining a data loading report based on the update process includes: The target database is determined based on domain characteristic data, and a database schema is generated based on the target database and the modified ontology structure. A graph framework is created based on the database schema. The triplet data is imported into the graph framework to obtain the import results. The import results are compared and verified. The database schema is updated based on the comparison and verification results. At the same time, the comparison and verification results and the import results are summarized, and a data loading report is generated using the update process.

[0059] In this embodiment, the target database refers to a specific database management system selected based on domain-specific data for storing and running the knowledge graph. The selection criteria include data scale, query patterns, and domain characteristics. For example, if the preset business domain is the semiconductor field, Neo4j graph database is selected for semiconductor patent graphs requiring complex relational reasoning; Apache Jena is selected for massive triple storage and SPARQL queries.

[0060] In this embodiment, the graph framework is an empty knowledge graph container with a predefined structure, created in the target database based on the database schema. It serves as the skeleton awaiting data import. For example, if the preset business domain is the semiconductor domain, the generated Cypher / SPARQL statements are executed in the selected graph database to create a blank graph containing all predefined node types, edge types, and constraints.

[0061] In this embodiment, the import result is a preliminary report generated after batch importing triplet data into the graph framework. It includes the number of successfully imported entries, entries that failed due to format or constraint conflicts, and the reasons for those failures. For example, if the preset business domain is semiconductor, the report shows that 98,500 triplets were successfully imported, and 15 failed due to excessively long manufacturer attribute values, with specific data for these failures listed.

[0062] The working principle and beneficial effects of the above embodiments are as follows: Based on domain characteristic data and the modified ontology structure, a database schema is generated. Based on this schema, a graph framework is created and triple data is imported. The import results are compared and verified. The data loading report is generated by combining the verification and import results, realizing automatic mapping from logical ontology to physical storage, improving the efficiency of knowledge graph construction, predicting data loading problems through comparison and verification, optimizing storage performance, and ensuring the quality of graph construction and system stability.

[0063] Example 5: The method provided in this embodiment of the present disclosure, wherein the parsing data loading report, based on the parsing results and combined with the modified ontology structure, determines the task execution sequence, and detects the execution process of the task execution sequence to complete the construction of the knowledge graph, includes: Parse the data loading report to obtain the data volume, task priority, and dependencies. Allocate computing resources based on the data volume, task priority, and dependencies, and generate a task execution sequence by modifying the ontology structure. The execution status and progress data of each task in the task execution sequence are collected in real time. The execution status and progress data are weighted by task, and the overall progress of knowledge graph construction is determined based on the task weighting results. Anomaly detection is performed based on the execution status and progress data, and a retry mechanism and anomaly prompt are triggered based on the anomaly detection results. Complete the construction of the knowledge graph.

[0064] In this embodiment, data volume, task priority, and dependencies refer to the scale of the triplet data to be processed. Task priority refers to the importance and urgency of different construction tasks. Dependencies refer to the sequential execution logic between tasks. For example, with 1 million data entries, the indexing task has a higher priority than data backup; the entity linking task can only begin after the entity identification task is completed.

[0065] In this embodiment, allocating computing resources is a process of dynamically allocating computing resources such as CPU, memory, and GPU to different tasks in the task execution sequence based on task parameters. For example, more CPU cores and memory are allocated to high-priority indexing tasks; multiple container instances are launched to process data cleaning tasks that can be parallelized simultaneously.

[0066] In this embodiment, the execution status and progress data are specific indicators collected in real time, reflecting the current running status of each task in the task execution sequence. For example, real-time monitoring data such as entity recognition task completed 75%, current CPU utilization at 85%, and relationship extraction task failing due to network timeout.

[0067] In this embodiment, the overall construction progress is a quantified percentage indicator reflecting the total progress of the knowledge graph construction, obtained by weighting the progress data of each task. For example, although the data cleaning task has been completed, the index building task, which has a higher weight, has just started. After weighted calculation, the system displays an overall construction progress of 30%.

[0068] In this embodiment, anomaly detection is a process of real-time analysis of execution status and progress data to automatically identify abnormal situations such as task failure, performance bottlenecks, or deviations from expectations. For example, if the system detects that the progress of a data import task has not changed within 10 minutes and memory usage continues to exceed 95%, it determines that the task has encountered an anomaly.

[0069] In this embodiment, the anomaly detection result is a specific conclusion output by the anomaly detection process, clearly indicating the task, type, and possible cause of the anomaly. For example, the detection result might be: the task 'Data Import' experienced a 'Memory Overflow' anomaly, caused by an excessively large batch of data being processed.

[0070] In this embodiment, the retry mechanism and the exception notification are as follows: The retry mechanism is the system's strategy to automatically re-execute failed tasks based on exception results. The exception notification is an alert message sent to the system administrator. For example, the system automatically halves the batch size of the data import task and re-executes it (retry mechanism), while simultaneously sending an email notification to the administrator: Task A has been automatically retried due to a memory exception, please pay attention (exception notification).

[0071] The working principle and beneficial effects of the above embodiments are as follows: parsing data to load reports to obtain task parameters, generating task sequences based on computing resource allocation and ontology structure correction, monitoring task status in real time and calculating overall progress with weights, triggering retry mechanisms through anomaly detection, and finally completing the knowledge graph construction. This achieves resource optimization and full-process monitoring of the construction process, improves task execution efficiency, ensures the stability of the construction process through dynamic anomaly handling, and ensures high-quality delivery of the knowledge graph.

[0072] To implement the above embodiments, this disclosure also proposes a knowledge graph ontology structure construction apparatus.

[0073] Figure 3 This is a schematic diagram of a knowledge graph ontology structure construction device provided in an embodiment of this disclosure. This device can be implemented by software and / or hardware, and is generally integrated into an electronic device. Figure 3As shown, the device includes: an interaction module, a construction module, a management module, and a detection module, wherein, The interaction module is used to extract domain characteristic data from a preset business domain, dynamically adjust the interaction model based on the domain characteristic data to obtain an adjusted interaction model, use the adjusted interaction model to extract the core entities and corresponding relationships between entities from the input data, and generate triplet data based on the core entities and corresponding relationships between entities. The construction module is used to cluster the core entities, generate a domain concept system, establish a mapping relationship between each core entity and the domain concept system, construct the ontology structure of the knowledge graph based on the domain concept system and the mapping relationship, and use the adjusted interaction model to detect and correct the ontology structure. The management module is used to generate a corresponding database schema based on the corrected ontology structure, compare and verify the triple data with the database schema, update the database schema based on the comparison and verification results, and obtain a data loading report based on the update process. The detection module is used to parse the data loading report, determine the task execution sequence based on the parsing results and the corrected ontology structure, and detect the execution process of the task execution sequence.

[0074] Figure 4 The knowledge graph automated construction architecture provided in this disclosure embodiment is for Figure 3 The extended diagram.

[0075] like Figure 4 As shown, in some possible implementations, the interaction module is specifically used for: The optimization unit is used to extract corresponding domain characteristic data from a preset business domain, dynamically adjust the prompt word template and inference parameters of the interaction model according to the domain characteristic data, and then optimize and adjust the interaction model to obtain the adjusted interaction model. The recognition unit is used to preprocess the input data, generate a standardized text dataset, and perform entity recognition on the standardized text dataset in combination with a preset business domain and an adjusted interaction model to obtain the core entities. The extraction unit is used to analyze the semantic associations between the core entities using the adjusted interaction model, extract the entity relationships between each core entity and other core entities based on the semantic associations, and generate triplet data based on the entity relationships and core entities.

[0076] like Figure 4 As shown, in some possible implementations, the building module is specifically used for: Clustering units are used to perform cluster analysis on the core entities, generate a domain concept system based on the cluster analysis results, and establish a mapping relationship between the domain concept system and the core entities. The ontology generation unit is used to dynamically generate the ontology structure of knowledge graphs based on mapping relationships and domain concept systems. The verification unit is used to verify whether there is a conflict in the ontology structure using a logical reasoning database. If there is a conflict, a correction scheme is generated using the adjusted interaction model, and the conflict is corrected using the correction scheme to obtain a corrected ontology structure.

[0077] like Figure 4 As shown, in some possible implementations, the management module is specifically used for: The adaptation unit is used to determine the target database based on domain characteristic data; The schema generation unit is used to generate a database schema based on the target database and the modified ontology structure. The report generation unit is used to create a graph framework based on the database schema, import the triplet data into the graph framework, obtain the import result, perform a comparison and verification on the import result, update the database schema based on the comparison and verification result, summarize the comparison and verification result and the import result, and generate a data loading report using the update process.

[0078] like Figure 4 As shown, in some possible implementations, the detection module includes: The sequence generation unit is used to parse the data loading report, obtain the data volume, task priority and dependency relationship, allocate computing resources based on the data volume, task priority and dependency relationship, and generate a task execution sequence in combination with the modified ontology structure; The progress determination unit is used to collect the execution status and progress data of each task in the task execution sequence in real time, perform task weighting on the execution status and progress data, and determine the overall progress of knowledge graph construction based on the task weighting results. An anomaly detection unit is used to perform anomaly detection based on the execution status and progress data, and trigger a retry mechanism and anomaly prompt based on the anomaly detection result. Complete the construction of the knowledge graph.

[0079] Figure 5 for Figure 1 , Figure 2 , Figure 3 as well as Figure 4 The timing flowchart during execution.

[0080] The knowledge graph ontology structure construction apparatus provided in this disclosure can execute the knowledge graph ontology structure construction method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

[0081] To implement the above embodiments, this disclosure also proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the knowledge graph ontology structure construction method described above.

[0082] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure.

[0083] The following is a detailed reference. Figure 6 The diagram illustrates a structural schematic suitable for implementing the electronic device 300 in the embodiments of this disclosure. The electronic device 300 in the embodiments of this disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0084] like Figure 6 As shown, the electronic device 300 may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a memory 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processor 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0085] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0086] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from memory 308, or installed from ROM 302. When the computer program is executed by processor 301, it performs the functions defined in the knowledge graph ontology structure construction method of embodiments of this disclosure.

[0087] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0088] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0089] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0090] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the aforementioned knowledge graph ontology structure construction method.

[0091] Electronic devices can be programmed with computer program code in one or more programming languages ​​or combinations thereof to perform the operations of this disclosure. These programming languages ​​include, but are not limited to, object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0092] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0093] The modules described in the embodiments of this disclosure can be implemented in software or hardware. The names of the modules are not, in some cases, intended to limit the functionality of the module itself.

[0094] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0095] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0096] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0097] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0098] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for constructing a knowledge graph ontology structure, characterized in that, include: Extract domain characteristic data from a preset business domain, adjust the interaction model based on the domain characteristic data to obtain an adjusted interaction model, use the adjusted interaction model to extract the core entities and corresponding relationships between entities from the input data, and generate triplet data based on the core entities and corresponding relationships between entities. The core entities are clustered to generate a domain concept system, and a mapping relationship between each core entity and the domain concept system is established. The ontology structure of the knowledge graph is constructed based on the domain concept system and the mapping relationship. The ontology structure is then detected and corrected using an adjusted interaction model. A corresponding database schema is generated based on the modified ontology structure. The database schema is compared and verified based on the triple data. The database schema is updated based on the comparison and verification results. A data loading report is obtained based on the update process. The data loading report is parsed, and the task execution sequence is determined based on the parsing results and the modified ontology structure. The execution process of the task execution sequence is then monitored to complete the construction of the knowledge graph.

2. The method according to claim 1, characterized in that, The process involves extracting domain characteristic data from a preset business domain, dynamically adjusting the interaction model based on the domain characteristic data to obtain an adjusted interaction model, using the adjusted interaction model to extract core entities and corresponding inter-entity relationships from the input data, and generating triplet data based on the core entities and corresponding inter-entity relationships, including: Extract corresponding domain characteristic data from the preset business domain, dynamically adjust the prompt word template and inference parameters of the interaction model based on the domain characteristic data, and then optimize and adjust the interaction model to obtain the adjusted interaction model. The input data is preprocessed to generate a standardized text dataset. The standardized text dataset is then combined with a preset business domain and an adjusted interaction model to perform entity recognition and obtain the core entities. The semantic relationships between the core entities are analyzed using the adjusted interaction model. Based on the semantic relationships, the entity relationships between each core entity and other core entities are extracted, and triplet data are generated based on the entity relationships and core entities.

3. The method according to claim 1, characterized in that, The process of clustering the core entities to generate a domain concept system and establishing a mapping relationship between each core entity and the domain concept system, constructing the ontology structure of the knowledge graph based on the domain concept system and the mapping relationship, and using an adjusted interaction model to detect and correct the ontology structure includes: Cluster analysis is performed on the core entities, and a domain concept system is generated based on the cluster analysis results. A mapping relationship between the domain concept system and the core entities is established. The ontology structure of knowledge graphs is dynamically generated based on mapping relationships and domain concept systems; The logical reasoning database is used to verify whether there are any conflicts in the ontology structure. If there are, a correction scheme is generated using the adjusted interaction model. The conflict is then corrected using the correction scheme to obtain the corrected ontology structure.

4. The method according to claim 1, characterized in that, The process involves generating a corresponding database schema based on the modified ontology structure, comparing and verifying the triplet data with the database schema, updating the database schema based on the comparison and verification results, and obtaining a data loading report based on the update process, including: The target database is determined based on domain characteristic data, and a database schema is generated based on the target database and the modified ontology structure. A graph framework is created based on the database schema. The triplet data is imported into the graph framework to obtain the import results. The import results are compared and verified. The database schema is updated based on the comparison and verification results. At the same time, the comparison and verification results and the import results are summarized, and a data loading report is generated using the update process.

5. The method according to claim 1, characterized in that, The parsed data loading report determines the task execution sequence based on the parsing results and the corrected ontology structure, and monitors the execution process of the task execution sequence to complete the construction of the knowledge graph, including: Parse the data loading report to obtain the data volume, task priority, and dependencies. Allocate computing resources based on the data volume, task priority, and dependencies, and generate a task execution sequence by modifying the ontology structure. The execution status and progress data of each task in the task execution sequence are collected in real time. The execution status and progress data are weighted by task, and the overall progress of knowledge graph construction is determined based on the task weighting results. Anomaly detection is performed based on the execution status and progress data, and a retry mechanism and anomaly prompt are triggered based on the anomaly detection results. Complete the construction of the knowledge graph.

6. A knowledge graph ontology structure construction device, the device comprising: The interaction module is used to extract domain characteristic data from a preset business domain, dynamically adjust the interaction model based on the domain characteristic data to obtain an adjusted interaction model, use the adjusted interaction model to extract the core entities and corresponding relationships between entities from the input data, and generate triplet data based on the core entities and corresponding relationships between entities. The construction module is used to cluster the core entities, generate a domain concept system, establish a mapping relationship between each core entity and the domain concept system, construct the ontology structure of the knowledge graph based on the domain concept system and the mapping relationship, and use the adjusted interaction model to detect and correct the ontology structure. The management module is used to generate a corresponding database schema based on the corrected ontology structure, compare and verify the triple data with the database schema, update the database schema based on the comparison and verification results, and obtain a data loading report based on the update process. The detection module is used to parse the data loading report, determine the task execution sequence based on the parsing results and the corrected ontology structure, and detect the execution process of the task execution sequence.

7. The apparatus according to claim 6, characterized in that, The interaction module includes: The optimization unit is used to extract corresponding domain characteristic data from a preset business domain, dynamically adjust the prompt word template and inference parameters of the interaction model according to the domain characteristic data, and then optimize and adjust the interaction model to obtain the adjusted interaction model. The recognition unit is used to preprocess the input data, generate a standardized text dataset, and perform entity recognition on the standardized text dataset in combination with a preset business domain and an adjusted interaction model to obtain the core entities. The extraction unit is used to analyze the semantic associations between the core entities using the adjusted interaction model, extract the entity relationships between each core entity and other core entities based on the semantic associations, and generate triplet data based on the entity relationships and core entities.

8. The apparatus according to claim 6, characterized in that, The building module includes: Clustering units are used to perform cluster analysis on the core entities, generate a domain concept system based on the cluster analysis results, and establish a mapping relationship between the domain concept system and the core entities. The ontology generation unit is used to dynamically generate the ontology structure of knowledge graphs based on mapping relationships and domain concept systems. The verification unit is used to verify whether there is a conflict in the ontology structure using a logical reasoning database. If there is a conflict, a correction scheme is generated using the adjusted interaction model, and the conflict is corrected using the correction scheme to obtain a corrected ontology structure.

9. An electronic device, characterized in that, include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the knowledge graph ontology structure construction method as described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, causes the processor to implement the knowledge graph ontology structure construction method provided in any one of claims 1-6.

Images