CN-122021826-A - Knowledge graph self-evolution method, device, equipment and storage medium

CN122021826ACN 122021826 ACN122021826 ACN 122021826ACN-122021826-A

Abstract

The application provides a knowledge graph self-evolution method, which is used for continuously updating an operation and maintenance knowledge graph of equipment to obtain the current latest operation and maintenance knowledge graph of the equipment, and comprises the steps of obtaining operation and maintenance data of the equipment and the current operation and maintenance knowledge graph of the equipment; extracting each triplet from the operation and maintenance data, carrying out quality evaluation on each triplet based on the current operation and maintenance knowledge graph, screening each triplet for updating the current operation and maintenance knowledge graph, merging each screened triplet into the current operation and maintenance knowledge graph to obtain an updated operation and maintenance knowledge graph, wherein the updated operation and maintenance knowledge graph is the current latest operation and maintenance knowledge graph of the equipment. The application updates the current operation and maintenance knowledge graph based on the acquired operation and maintenance data, can capture new operation and maintenance data generated in the process of identifying operation and maintenance in real time, and can realize continuous update of the equipment operation and maintenance knowledge graph.

Inventors

ZHANG CAIFENG
ZHENG ZIHAO

Assignees

北京至简能源有限公司

Dates

Publication Date: 20260512
Application Date: 20260120

Claims (10)

1. The knowledge graph self-evolution method is characterized by being used for continuously updating the operation and maintenance knowledge graph of equipment to obtain the current latest operation and maintenance knowledge graph of the equipment, and comprising the following steps of: acquiring operation and maintenance data of the equipment and a current operation and maintenance knowledge graph; Extracting each triplet from the operation and maintenance data, carrying out quality evaluation on each triplet based on the current operation and maintenance knowledge graph, and screening out each triplet for updating the current operation and maintenance knowledge graph; and merging each screened triplet into the current operation and maintenance knowledge graph to obtain an updated operation and maintenance knowledge graph, wherein the updated operation and maintenance knowledge graph is the current latest operation and maintenance knowledge graph of the equipment.
2. The knowledge graph self-evolution method according to claim 1, further comprising the operation of storing the operation data according to a preset storage format after the operation data of the equipment is obtained; The preset storage format is device type-log type-time stamp.
3. The method of claim 1, wherein the dimension of quality assessment includes at least one of consistency, integrity, and reliability when quality assessment is performed on each of the triples based on the current operation and maintenance knowledge graph.
4. A knowledge-graph self-evolution method according to claim 3, wherein, in said evaluating the integrity of each of said triples, it comprises: when the data of each triplet is confirmed to be complete, entering the consistency evaluation stage for each triplet; and when confirming that the triples with incomplete data exist in the triples, entering a manual review stage by the triples with incomplete data.
5. The method according to claim 4, wherein when performing the consistency assessment for each triplet with complete data, the method comprises: acquiring the triplet with complete data and the current operation and maintenance knowledge graph; Screening out all the triples inconsistent with the triples in the current operation and maintenance knowledge graph from all the triples with complete data based on the current operation and maintenance knowledge graph, and obtaining the sources of the screened triples; And further screening out reserved triples from the screened triples according to the sources, wherein the reserved triples enter a reliability evaluation stage.
6. The method according to claim 5, wherein the reliability evaluation of each of the remaining triples comprises: Calculating the credibility score of each reserved triplet; and screening out the triples with the reliability score larger than a preset threshold value from the reliability scores, wherein the screened triples are used as the triples for updating the current operation and maintenance knowledge graph, which are obtained after quality evaluation.
7. The knowledge-graph self-evolution method according to claim 6, wherein the calculation of the confidence score of each of the remaining triples is performed by calculating at least one of a source score and a representation clarity score.
8. The utility model provides a knowledge graph self-evolution device which is characterized in that, be used for the operation and maintenance knowledge graph to equipment constantly update and obtain the current latest operation and maintenance knowledge graph of equipment, include: The data acquisition module is used for acquiring operation and maintenance data of the equipment and a current operation and maintenance knowledge graph; The quality evaluation module is used for extracting each triplet from the operation and maintenance data, evaluating the quality of each triplet based on the current operation and maintenance knowledge graph, and screening out each triplet used for updating the current operation and maintenance knowledge graph; And the knowledge map updating module is used for merging each screened triplet into the current operation and maintenance knowledge map to obtain an updated operation and maintenance knowledge map, wherein the updated operation and maintenance knowledge map is the current latest operation and maintenance knowledge map of the equipment.
9. A license verifying apparatus, characterized by comprising: A processor; A memory for storing processor-executable instructions; wherein the processor is configured to implement the method of any one of claims 1 to 7 when executing the executable instructions.
10. A non-transitory computer readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method of any of claims 1 to 7.

Description

Knowledge graph self-evolution method, device, equipment and storage medium Technical Field The present application relates to the field of knowledge graph update technologies, and in particular, to a method, an apparatus, a device, and a storage medium for knowledge graph self-evolution. Background The operation and maintenance of the equipment at present are usually carried out depending on the operation and maintenance knowledge graph of the equipment, but the operation and maintenance knowledge graph of the equipment is required to be continuously updated according to the actual operation and maintenance scene, the current updating mode usually adopts a direct real-time operation and maintenance interaction log (such as dialogue record) as an unstructured text base, so that the current operation and maintenance answer is searched from the unstructured text base through a search or simple question-answer model, and the operation and maintenance knowledge graph is updated by carrying out one-time or periodic batch knowledge extraction through an NLP technology through a manual mobile phone history document (such as a maintenance report and a manual) at all times, so that a new operation and maintenance knowledge graph of the equipment is obtained. In the updating mode, the operation and maintenance knowledge graph is static in nature, and can not capture real-time interaction to generate new operation and maintenance knowledge in time during updating, so that the problem of slow updating exists. Therefore, how to make the operation and maintenance knowledge graph of the equipment continuously self-evolve is a problem to be solved by the person skilled in the art. Disclosure of Invention In view of this, the application provides a knowledge graph self-evolution method, a device, equipment and a storage medium, which can realize continuous self-evolution of the equipment operation and maintenance knowledge graph. According to a first aspect of the present application, there is provided a knowledge graph self-evolution method for continuously updating an operation and maintenance knowledge graph of a device to obtain a current latest operation and maintenance knowledge graph of the device, the method comprising: acquiring operation and maintenance data of the equipment and a current operation and maintenance knowledge graph; Extracting each triplet from the operation and maintenance data, carrying out quality evaluation on each triplet based on the current operation and maintenance knowledge graph, and screening out each triplet for updating the current operation and maintenance knowledge graph; and merging each screened triplet into the current operation and maintenance knowledge graph to obtain an updated operation and maintenance knowledge graph, wherein the updated operation and maintenance knowledge graph is the current latest operation and maintenance knowledge graph of the equipment. In one possible implementation manner, after the operation data of the device is obtained, the operation of storing the operation data according to a preset storage format is further included; The preset storage format is device type-log type-time stamp. In one possible implementation, when performing quality assessment of each of the triples based on the current operation and maintenance knowledge graph, the dimension in which quality assessment is performed includes at least one of consistency, integrity, and reliability. In one possible implementation, when performing the integrity assessment for each of the triples, the method includes: when the data of each triplet is confirmed to be complete, entering the consistency evaluation stage for each triplet; and when confirming that the triples with incomplete data exist in the triples, entering a manual review stage by the triples with incomplete data. In one possible implementation, when performing the consistency assessment for each of the triples for which data is complete, the method includes: acquiring the triplet with complete data and the current operation and maintenance knowledge graph; Screening out all the triples inconsistent with the triples in the current operation and maintenance knowledge graph from all the triples with complete data based on the current operation and maintenance knowledge graph, and obtaining the sources of the screened triples; And further screening out reserved triples from the screened triples according to the sources, wherein the reserved triples enter a reliability evaluation stage. In one possible implementation, when evaluating the reliability of each reserved triplet, it includes: Calculating the credibility score of each reserved triplet; and screening out the triples with the reliability score larger than a preset threshold value from the reliability scores, wherein the screened triples are used as the triples for updating the current operation and maintenance knowledge graph, which are obtained after quality evaluation. In one