A method and device for constructing an operation and maintenance knowledge graph, and an electronic device

Abstract

The application discloses a construction method and device of an operation and maintenance knowledge graph and electronic equipment, and the method comprises the following steps: obtaining domain knowledge content of a target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is a domain to which an operation and maintenance system belongs; constructing domain graph data according to the domain knowledge content, wherein the domain graph data comprises a plurality of first nodes, and any two first nodes are directly connected or indirectly connected through other first nodes; constructing case graph data according to the operation and maintenance case content, wherein the case graph data is composed of at least one subgraph data, the subgraph data are connected through a center node, the subgraph data comprise a plurality of second nodes, and any two second nodes are directly connected or indirectly connected through other second nodes; and performing fusion processing on the domain graph data and the case graph data to obtain an operation and maintenance knowledge graph corresponding to the target domain.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus and electronic device for constructing an operation and maintenance knowledge graph. Background Technology

[0002] As the number of business systems in enterprises increases daily, the hardware and software architectures of these systems become increasingly complex, leading to more intricate maintenance and operation tasks. With system architecture upgrades, updated technology selections, and iterative equipment or software versions, the skills and knowledge that maintenance personnel need to master are becoming increasingly diverse.

[0003] In today's operations and maintenance work, the operation and maintenance personnel of different system modules mostly communicate and exchange relevant professional knowledge through verbal explanations, technical documents, or experience sharing. However, the relevant knowledge service platforms are limited by web page storage and search engines, making it difficult to provide accurate search and query services for technical personnel in need.

[0004] Therefore, there is an urgent need for a technical solution that can build a knowledge graph of operation and maintenance knowledge to facilitate retrieval. Summary of the Invention

[0005] In view of this, this application provides a method, apparatus, medium, and electronic device for constructing an operations and maintenance knowledge graph, used to construct an operations and maintenance knowledge graph to facilitate the querying of operations and maintenance knowledge. As follows:

[0006] A method for constructing an operations and maintenance knowledge graph, the method comprising:

[0007] Obtain domain knowledge content of the target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs;

[0008] Based on the domain knowledge content, a domain graph data is constructed. The domain graph data contains multiple first nodes. Any two first nodes are directly connected or indirectly connected through other first nodes. The first node represents the domain knowledge content.

[0009] Based on the operation and maintenance case content, case graph data is constructed. The case graph data consists of at least one sub-graph data. The sub-graph data are connected through a central node. The sub-graph data contains multiple second nodes. Any two second nodes are directly connected or indirectly connected through other second nodes. The second node represents the operation and maintenance case content to which it belongs.

[0010] The domain graph data and the case graph data are fused to obtain the operation and maintenance knowledge graph corresponding to the target domain.

[0011] The above method, preferably, involves constructing domain graph data based on the domain knowledge content, including:

[0012] Based on a pre-set ontology model architecture, the domain knowledge content is processed to obtain a domain ontology model corresponding to the domain knowledge content. The domain ontology model contains multiple domain ontology.

[0013] According to the preset mapping rules, the domain ontology is converted into a first node based on graph data to obtain domain graph data.

[0014] In the above method, preferably, the mapping rule includes: the conversion relationship between ontology constituent elements in the ontology and node storage elements in the node;

[0015] Specifically, according to a preset mapping rule, the domain ontology is converted into a first node based on graph data to obtain domain graph data, including:

[0016] According to the transformation relationship, the ontology components in the domain ontology are transformed into node storage elements in the graph data nodes to obtain the first node based on the graph data, and the first node constitutes the domain graph data.

[0017] In the above method, preferably, each of the above operation and maintenance case contents corresponds to one operation and maintenance event;

[0018] The step of constructing case graph data based on the content of the operation and maintenance case includes:

[0019] Based on the operation and maintenance case content corresponding to the operation and maintenance event, construct subgraph data corresponding to the operation and maintenance event; each subgraph data corresponding to the operation and maintenance event has a root node, and the second node in each subgraph data is directly or indirectly connected to the root node;

[0020] The root nodes of all the subgraph data are concatenated through the central node to obtain the case graph data.

[0021] In the above method, preferably, the operation and maintenance case content includes the table record content of the corresponding operation and maintenance event, and the table record content consists of the content of multiple cells;

[0022] The step of constructing the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event includes:

[0023] Using the cells in the table records as nodes of graph data, subgraph data corresponding to the operation and maintenance event is constructed based on the content of the cells.

[0024] In the above method, preferably, the operation and maintenance case content includes the document record content of the corresponding operation and maintenance event, and the document record content includes at least one record statement;

[0025] The step of constructing the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event includes:

[0026] The record statements corresponding to the operation and maintenance events are extracted into triples to obtain multiple entities and the relationships between the entities;

[0027] Based on the entities and the relationships between them, construct the subgraph data corresponding to the operation and maintenance events.

[0028] In the above method, preferably, the operation and maintenance case content includes table records and document records of the corresponding operation and maintenance events;

[0029] The step of constructing the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event includes:

[0030] Based on the table record content corresponding to the operation and maintenance event, construct the first graph data corresponding to the operation and maintenance event;

[0031] Based on the document record content corresponding to the operation and maintenance event, construct the second graph data corresponding to the operation and maintenance event;

[0032] Align the nodes in the first graph data and the nodes in the second graph data in terms of semantic description;

[0033] The first graph data and the second graph data are merged to obtain the subgraph data corresponding to the operation and maintenance event.

[0034] Preferably, in the above method, the step of fusing the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain includes:

[0035] Align the nodes in the domain graph data and the nodes in the case graph data in terms of semantic description;

[0036] Establish a parent-child relationship between at least one node in the case graph data and a node in the domain graph data, so that at least one node in the case graph data has a parent node in the domain graph data, to obtain the operation and maintenance knowledge graph corresponding to the target domain;

[0037] Among them, the node storage elements of the two nodes that have established the parent-child relationship satisfy the node association condition.

[0038] An apparatus for constructing an operations and maintenance knowledge graph, the apparatus comprising:

[0039] The content acquisition unit is used to acquire domain knowledge content of the target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs;

[0040] A domain graph construction unit is used to construct domain graph data based on the domain knowledge content. The domain graph data includes multiple first nodes, and any first node has a connection relationship with at least one other first node. The first node represents the domain knowledge content.

[0041] The case diagram construction unit is used to construct case diagram data based on the operation and maintenance case content. The case diagram data consists of at least one sub-diagram data, which are connected to each other through a central node. The sub-diagram data contains multiple second nodes, and any second node has a connection relationship with at least one other second node. The second node represents the operation and maintenance case content to which it belongs.

[0042] The graph fusion unit is used to perform graph data fusion processing on the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain.

[0043] An electronic device, comprising:

[0044] A memory used to store computer programs and the data generated by the execution of said computer programs;

[0045] A processor is configured to execute the computer program to: obtain domain knowledge content of a target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs; construct domain graph data based on the domain knowledge content, wherein the domain graph data contains multiple first nodes, any first node having a connection relationship with at least one other first node, and the first node representing the domain knowledge content; construct case graph data based on the operation and maintenance case content, wherein the case graph data consists of at least one subgraph data, the subgraph data being connected through a central node, the subgraph data containing multiple second nodes, any second node having a connection relationship with at least one other second node, and the second node representing the operation and maintenance case content to which it belongs; and perform graph data fusion processing on the domain graph data and the case graph data to obtain an operation and maintenance knowledge graph corresponding to the target domain.

[0046] As can be seen from the above technical solutions, the method, apparatus, and electronic device for constructing an operation and maintenance knowledge graph disclosed in this application, for the target domain to which the operation and maintenance system belongs, acquires the domain knowledge content and at least one operation and maintenance case content of the target domain, and then constructs domain graph data and case graph data respectively. Then, the domain graph data and case graph data are fused to obtain the operation and maintenance knowledge graph of the domain to which the operation and maintenance system belongs. It is evident that this application, by constructing graph data of domain knowledge content and operation and maintenance case content separately and then fusing the graph data, obtains a more comprehensive and complete operation and maintenance knowledge graph represented by graph data within the corresponding domain, facilitating the query of operation and maintenance knowledge. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 A flowchart illustrating a method for constructing an operation and maintenance knowledge graph, as provided in Embodiment 1 of this application;

[0049] Figure 2 This is an example diagram of the domain graph data constructed in the embodiments of this application;

[0050] Figure 3 This is an example diagram of the case diagram data constructed in the embodiments of this application;

[0051] Figure 4 A partial flowchart of a method for constructing an operation and maintenance knowledge graph provided in Embodiment 1 of this application;

[0052] Figure 5 This is another part of the flowchart of a method for constructing an operation and maintenance knowledge graph provided in Embodiment 1 of this application;

[0053] Figure 6 This is a schematic diagram of the structure of an operation and maintenance knowledge graph construction device provided in Embodiment 2 of this application;

[0054] Figure 7 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of this application;

[0055] Figure 8 Example diagram for constructing an operation and maintenance knowledge graph applicable to the business systems of financial enterprises in this application;

[0056] Figure 9 This is an example diagram illustrating a case knowledge graph for constructing documented records in the operation and maintenance knowledge graph construction of a financial enterprise business system, applicable to this application. Detailed Implementation

[0057] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0058] refer to Figure 1 This is a flowchart illustrating the implementation of a method for constructing an operations and maintenance knowledge graph according to Embodiment 1 of this application. This method is applicable to electronic devices capable of data processing, such as computers or servers. The technical solution in this embodiment is mainly used to construct an operations and maintenance knowledge graph belonging to the domain of the operations and maintenance system, facilitating the retrieval of operations and maintenance knowledge.

[0059] Specifically, the method in this embodiment may include the following steps:

[0060] Step 101: Obtain domain knowledge content for the target domain and at least one operation and maintenance case content for the target domain.

[0061] The target domain refers to the domain to which the operation and maintenance system belongs, such as the mobile phone domain, substation domain, or medical device domain. The domain knowledge content of the target domain refers to the inherent knowledge domain of the target domain, that is, storing definite and infrequently updated domain knowledge. This type of knowledge content mainly includes: system operation and maintenance knowledge, upstream and downstream system knowledge, etc., and is mainly acquired from: industry standards and specifications, software technical architecture, development documents, etc. Taking the mobile phone domain as an example, the domain knowledge content of the mobile phone domain may include: mobile phone industry standards and specifications, mobile phone software technical architecture, mobile phone hardware and software development documents, etc. The operation and maintenance case content of the target domain is the instance data domain. This type of knowledge content mainly includes the content of historical operation and maintenance cases, that is, storing some descriptive, real events. The source of this knowledge content is mainly the document records and / or table records left by historical operation and maintenance activities, such as the table records left by the repair of a specific mobile phone model.

[0062] Step 102: Construct domain graph data based on domain knowledge content.

[0063] Domain graph data refers to graph data corresponding to domain knowledge content. Domain graph data contains multiple first nodes, and each first node has a connection relationship with at least one other first node, such as... Figure 2As shown, any two first nodes are directly connected or indirectly connected through other first nodes. In other words, any first node in the domain graph data can be connected to any other first node through one or more first nodes. The first nodes and their corresponding connections can represent the domain knowledge content.

[0064] In one implementation, this embodiment can extract triples from domain knowledge content, such as identifying entities and relationships between entities, and then construct domain graph data based on the extracted triples.

[0065] Step 103: Construct case diagram data based on the content of the operation and maintenance case.

[0066] Case study graph data refers to the graph data corresponding to the content of operation and maintenance cases. Case study graph data consists of at least one sub-graph data, which are connected through a central node. Each sub-graph data may contain multiple second nodes, and any second node is connected to at least one other second node, such as... Figure 3 As shown, any two second nodes are directly connected or indirectly connected through other second nodes. In other words, within each subgraph, any second node can be connected to any other second node through one or more other second nodes. These second nodes and their corresponding connections characterize the content of the operational case.

[0067] It should be noted that the execution order between steps 102 and 103 is not limited to... Figure 1 The execution order can be either step 103 first and then step 102, or steps 102 and 103 can be executed simultaneously. All the different technical solutions formed are within the protection scope of this application.

[0068] Step 104: Perform graph data fusion processing on the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain.

[0069] The operation and maintenance knowledge graph corresponding to the target domain can be understood as graph data corresponding to the target domain. This operation and maintenance knowledge graph consists of domain graph data and case graph data.

[0070] Specifically, in step 104, graph data fusion can be achieved in the following way to obtain the operation and maintenance knowledge graph:

[0071] First, align the nodes in the domain graph data and the nodes in the case graph data in terms of semantic description. For example, unify the storage elements such as node name, node attributes, and node relationships between the first and second nodes, so that the node names corresponding to the same node are consistent, the attribute names corresponding to the same node attributes are consistent, and so on.

[0072] Next, a parent-child relationship is established between at least one node in the case study graph data and a node in the domain graph data, so that at least one node in the case study graph data has a parent node in the domain graph data, thereby obtaining the operation and maintenance knowledge graph corresponding to the target domain. For example, for each first node in the case study graph data, a second node is found in the domain graph data to establish a parent-child relationship between the first and second nodes, with the second node as the parent node and the first node as the child node. Thus, the fusion of the domain graph data and the case study graph data is achieved by establishing parent-child relationships between nodes.

[0073] Specifically, the node association conditions must be met in the node storage elements of the two nodes that have a parent-child relationship. For example, the similarity of the descriptions of the first and second nodes that have a parent-child relationship in terms of node attributes and other elements must be greater than or equal to a similarity threshold, and / or, the first and second nodes that have a parent-child relationship must have a logical relationship in terms of node relationships and other elements, such as causal logical relationship, sequential logical relationship, parallel logical relationship, or hierarchical logical relationship.

[0074] As can be seen from the above technical solutions, in the method for constructing an operation and maintenance knowledge graph provided in Embodiment 1 of this application, for the target domain to which the operation and maintenance system belongs, domain knowledge content and at least one operation and maintenance case content of the target domain are obtained respectively, and then domain graph data and case graph data are constructed respectively. Then, the domain graph data and case graph data are fused to obtain the operation and maintenance knowledge graph of the domain to which the operation and maintenance system belongs. It is evident that in this embodiment, by constructing graph data of domain knowledge content and operation and maintenance case content separately and then fusing the graph data, a more comprehensive and complete operation and maintenance knowledge graph represented by graph data in the corresponding domain is obtained, facilitating the query of operation and maintenance knowledge.

[0075] In one implementation, step 101, when constructing domain graph data based on domain knowledge content, can be achieved through the following steps, such as... Figure 4 As shown:

[0076] Step 401: Process the domain knowledge content according to the pre-set ontology model architecture to obtain the domain ontology model corresponding to the domain knowledge content.

[0077] The domain ontology model contains multiple domain ontology elements. A domain ontology includes ontology components such as attributes, relations, classes, instances, functions, and constraints.

[0078] Specifically, the ontology model architecture can be pre-configured by staff according to the target domain, such as star architecture, ring architecture, or tree architecture, to represent the compositional architecture between domain ontology.

[0079] Based on this, in this embodiment, the data in the domain knowledge content is parsed and extracted according to the preset ontology model architecture, and then a domain ontology model under the ontology model architecture is constructed. The composition architecture between domain ontology in the domain ontology model matches the ontology model architecture, and the domain ontology in the domain ontology model and the relationship between domain ontology represent the domain knowledge content.

[0080] For example, in this embodiment, a corresponding ontology model building tool such as Protégé can be used to process the domain knowledge content under the ontology model architecture, thereby obtaining the domain ontology model corresponding to the domain knowledge content.

[0081] Step 402: According to the preset mapping rules, the domain ontology model is converted into the first node based on graph data to obtain the domain graph data.

[0082] Here, the mapping rule refers to the rule for converting ontology into nodes based on graph data. Based on this, in this embodiment, the domain ontology in the domain ontology model is converted into the first node based on graph data according to the mapping rule, thereby obtaining the domain graph data composed of the first node.

[0083] Specifically, the mapping rules can include the transformation relationship between ontology components in the ontology and node storage elements in the nodes. Based on this, step 402 can be implemented in the following way:

[0084] According to the transformation relationship, the ontology components in the domain ontology are transformed into the node storage elements in the corresponding graph data nodes, so that the domain ontology is transformed into graph data nodes, thereby obtaining the first node based on graph data. The first node constitutes the domain graph data.

[0085] For example, the mapping rules can be shown in Table 1:

[0086] Table 1 Mapping Rules

[0087]

[0088]

[0089] Based on this, in this embodiment, the domain ontology is converted into graph data nodes according to the above mapping relationship to obtain domain graph data.

[0090] In one implementation, each operation and maintenance case corresponds to an operation and maintenance event, which can be identified using an event identifier such as an event ID. Based on this, step 103, when constructing the case graph data according to the operation and maintenance case content, can be implemented through the following steps, such as... Figure 5 As shown:

[0091] Step 501: Based on the operation and maintenance case content corresponding to the operation and maintenance event, construct the subgraph data corresponding to the operation and maintenance event.

[0092] Each operation and maintenance event corresponds to a subgraph data with a root node, and the second node in each subgraph data is directly or indirectly connected to the root node.

[0093] In other words, in this embodiment, the corresponding operation and maintenance case content can be processed for each operation and maintenance event to construct subgraph data corresponding to each operation and maintenance event. The number of operation and maintenance events determines the number of subgraph data.

[0094] Step 502: Concatenate the root nodes of all subgraph data through the center node to obtain the case graph data.

[0095] In this embodiment, the central node can be a node representing the target domain. Based on this, the root node of each subgraph data is connected to the central node, thereby obtaining complete case graph data.

[0096] In practice, the content of each operation and maintenance event may be tabular records, document records, or both. Tabular records consist of the contents of multiple cells, while document records contain at least one record statement.

[0097] In one implementation, when the operation and maintenance case content corresponding to the operation and maintenance event only contains table records, this embodiment constructs the subgraph data corresponding to the operation and maintenance event by using the cells in the table records as nodes based on the graph data. The subgraph data is then constructed based on the content of the cells. For example, the second node contained in the subgraph data is first initialized according to the cells in the table records, and then the node storage elements of the second node are set according to the content of the cells, thereby obtaining complete and accurate subgraph data.

[0098] In one implementation, when the operation and maintenance case content corresponding to the operation and maintenance event only contains document record content, this embodiment extracts triples from the record statements corresponding to the operation and maintenance event when constructing the subgraph data corresponding to the operation and maintenance event to obtain multiple entities and the relationships between these entities; then, based on the entities and the relationships between the entities, the subgraph data corresponding to the operation and maintenance event is constructed.

[0099] Specifically, in this embodiment, the document record content can first be converted into a text document format, then the text document can be serialized into words, then the serialized words can be identified into entities. After the entities are identified and extracted, the corresponding relationships are extracted, and then the second node in the subgraph data can be constructed based on the entities. Based on the relationships between entities, the node storage elements of the second node can be constructed, thereby obtaining complete and accurate subgraph data.

[0100] In one implementation, when the operation and maintenance case content corresponding to the operation and maintenance event includes both table records and document records, in this embodiment, when constructing the subgraph data corresponding to the operation and maintenance event, the first graph data corresponding to the operation and maintenance event can be constructed firstly based on the table records corresponding to the operation and maintenance event. The method for constructing the first graph data can refer to the relevant content above. Furthermore, the second graph data corresponding to the operation and maintenance event can be constructed based on the document records corresponding to the operation and maintenance event. The method for constructing the second graph data can refer to the relevant content above. Based on this, the nodes in the first graph data and the nodes in the second graph data are aligned semantically. The alignment method can refer to the relevant content above. Finally, the first graph data and the second graph data are merged. For example, using any graph data from the first graph data and the second graph data as a benchmark, the benchmark graph data is optimized using the other graph data, thereby obtaining complete and accurate subgraph data.

[0101] refer to Figure 6 This is a schematic diagram of a device for constructing an operation and maintenance knowledge graph according to Embodiment 2 of this application. This device can be applied to electronic devices capable of data processing, such as computers or servers. The technical solution in this embodiment is mainly used to construct an operation and maintenance knowledge graph belonging to the domain of the operation and maintenance system, so as to facilitate operation and maintenance knowledge retrieval.

[0102] Specifically, the device in this embodiment may include the following units:

[0103] The content acquisition unit 601 is used to acquire domain knowledge content of the target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs;

[0104] The domain graph construction unit 602 is used to construct domain graph data based on the domain knowledge content. The domain graph data includes multiple first nodes, and any first node has a connection relationship with at least one other first node. The first node represents the domain knowledge content.

[0105] The case diagram construction unit 603 is used to construct case diagram data based on the operation and maintenance case content. The case diagram data consists of at least one sub-diagram data, which are connected to each other through a central node. The sub-diagram data contains multiple second nodes, and any second node has a connection relationship with at least one other second node. The second node represents the operation and maintenance case content to which it belongs.

[0106] The graph fusion unit 604 is used to perform graph data fusion processing on the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain.

[0107] As can be seen from the above technical solution, in the operation and maintenance knowledge graph construction device provided in Embodiment 2 of this application, for the target domain to which the operation and maintenance system belongs, domain knowledge content and at least one operation and maintenance case content of the target domain are obtained respectively, and then domain graph data and case graph data are constructed respectively. Then, the domain graph data and case graph data are fused to obtain the operation and maintenance knowledge graph of the domain to which the operation and maintenance system belongs. It is evident that in this embodiment, by constructing graph data of domain knowledge content and operation and maintenance case content separately and then fusing the graph data, a more comprehensive and complete operation and maintenance knowledge graph represented by graph data in the corresponding domain is obtained, facilitating the query of operation and maintenance knowledge.

[0108] In one implementation, the domain graph construction unit 602 is specifically used to: process the domain knowledge content according to a pre-set ontology model architecture to obtain a domain ontology model corresponding to the domain knowledge content, wherein the domain ontology model contains multiple domain ontology; and convert the domain ontology into a first node based on graph data according to a preset mapping rule to obtain domain graph data.

[0109] In the optional implementation, the mapping rules include: the conversion relationship between the ontology constituent elements in the ontology and the node storage elements in the node;

[0110] Specifically, when the domain graph construction unit 602 converts the domain ontology into a first node based on graph data, it can do so in the following way: according to the conversion relationship, the ontology constituent elements in the domain ontology are converted into node storage elements in the graph data nodes to obtain the first node based on graph data, and the first node constitutes the domain graph data.

[0111] In one implementation, each operation and maintenance case content corresponds to an operation and maintenance event; wherein, the case graph construction unit 603 is specifically used to: construct subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event; each subgraph data corresponding to the operation and maintenance event has a root node, and the second node in each subgraph data is directly or indirectly connected to the root node; and concatenate the root nodes of all the subgraph data through the central node to obtain the case graph data.

[0112] In an optional implementation, the operation and maintenance case content includes table records of corresponding operation and maintenance events, and the table records consist of the contents of multiple cells;

[0113] Specifically, when the case graph construction unit 603 constructs the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event, it can do so in the following way: taking the cell in the table record content as the node based on the graph data, and constructing the subgraph data corresponding to the operation and maintenance event based on the content of the cell.

[0114] In an optional implementation, the operation and maintenance case content includes the document record content of the corresponding operation and maintenance event, and the document record content includes at least one record statement;

[0115] Specifically, when the case graph construction unit 603 constructs the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event, it can do so in the following way: extract triples from the record statements corresponding to the operation and maintenance event to obtain multiple entities and the relationships between the entities; construct the subgraph data corresponding to the operation and maintenance event based on the entities and the relationships between the entities.

[0116] In an optional implementation, the operation and maintenance case content includes table records and document records of the corresponding operation and maintenance events;

[0117] Specifically, when the case graph construction unit 603 constructs the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content, it can do so in the following ways: construct the first graph data corresponding to the operation and maintenance event based on the table record content corresponding to the operation and maintenance event; construct the second graph data corresponding to the operation and maintenance event based on the document record content corresponding to the operation and maintenance event; align the nodes in the first graph data and the nodes in the second graph data in terms of semantic description; and merge the first graph data and the second graph data to obtain the subgraph data corresponding to the operation and maintenance event.

[0118] In one implementation, the graph fusion unit 604 is specifically used to: align the nodes in the domain graph data and the nodes in the case graph data in terms of semantic description; establish a parent-child relationship between at least one node in the case graph data and a node in the domain graph data, such that at least one node in the case graph data has a parent node in the domain graph data, so as to obtain the operation and maintenance knowledge graph corresponding to the target domain; wherein, the node storage elements of the two nodes that have established the parent-child relationship satisfy the node association condition.

[0119] It should be noted that the specific implementation of each unit in this embodiment can be referred to the corresponding content above, and will not be described in detail here.

[0120] refer to Figure 7 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of this application. The electronic device is a computer or server capable of data processing. Specifically, the electronic device may include the following structure:

[0121] Memory 701 is used to store computer programs and data generated by the execution of said computer programs;

[0122] Processor 702 is configured to execute the computer program to: obtain domain knowledge content of a target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs; construct domain graph data based on the domain knowledge content, wherein the domain graph data contains multiple first nodes, any first node having a connection relationship with at least one other first node, and the first node representing the domain knowledge content; construct case graph data based on the operation and maintenance case content, wherein the case graph data consists of at least one subgraph data, the subgraph data being connected through a central node, the subgraph data containing multiple second nodes, any second node having a connection relationship with at least one other second node, and the second node representing the operation and maintenance case content to which it belongs; and perform graph data fusion processing on the domain graph data and the case graph data to obtain an operation and maintenance knowledge graph corresponding to the target domain.

[0123] As can be seen from the above technical solutions, in the electronic device provided in Embodiment 3 of this application, for the target domain to which the operation and maintenance system belongs, domain knowledge content and at least one operation and maintenance case content of the target domain are obtained respectively, and then domain graph data and case graph data are constructed respectively. Then, the domain graph data and case graph data are fused to obtain an operation and maintenance knowledge graph of the domain to which the operation and maintenance system belongs. It is evident that in this embodiment, by constructing graph data of domain knowledge content and operation and maintenance case content separately and then fusing the graph data, a more comprehensive and complete operation and maintenance knowledge graph represented by graph data in the corresponding domain is obtained, facilitating the query of operation and maintenance knowledge.

[0124] This application also provides a computer-readable medium storing a computer program, characterized in that the computer program, when run by a processor, implements the method for constructing an operation and maintenance knowledge graph as described in the preceding embodiments.

[0125] This application also provides a processor for running a program, wherein the program implements the method for constructing the operation and maintenance knowledge graph described in the above method embodiments.

[0126] This application also provides a computer program product that, when executed on a data processing device, enables the data processing device to implement the operation and maintenance knowledge graph construction method described in the above method embodiments.

[0127] Taking the operation and maintenance system (O&M) of a financial enterprise as an example, this application addresses the problems of complex and insufficiently structured knowledge systems in current system O&M work by constructing a knowledge graph for the O&M system, based on the technical solutions described above. It employs graph data as a key technology to summarize, express, and store the O&M knowledge (including domain knowledge and O&M cases). The basic solution involves ontology modeling, element mapping, and graph data embedding of the business system's O&M domain knowledge (i.e., the domain knowledge content mentioned above). It also involves triple extraction and graph data embedding of the business system's O&M case knowledge (i.e., the O&M case content mentioned above). After graph data fusion, the O&M knowledge base is constructed, providing a more convenient and efficient knowledge retrieval interface for technical and O&M personnel.

[0128] Specifically, this application proposes a knowledge graph construction scheme based on graph databases, using knowledge graphs for knowledge base modeling. The operation and maintenance knowledge graph of this application is divided into two parts: a domain knowledge graph (i.e., the domain graph data mentioned above) and a case knowledge graph (i.e., the case graph data mentioned above). The domain knowledge graph is an inherent knowledge domain, storing definite and infrequently updated domain knowledge. This type of knowledge mainly includes system operation and maintenance knowledge, upstream and downstream system knowledge, etc., and is primarily acquired from industry standards and specifications, software technical architecture, development documents, etc. The case knowledge graph is an instance data domain, mainly including historical operation and maintenance cases, storing facts and real events that have occurred. This type of knowledge primarily comes from textual and tabular records left from historical operation and maintenance activities.

[0129] From the perspective of knowledge graphs, domain knowledge graphs mostly exist in the form of classes, representing concepts abstracted from concrete things, while case knowledge graphs mostly exist in the form of instances, meaning each node represents only one thing. Due to the different meanings and sources of their stored information, their construction methods and processes are described below:

[0130] The knowledge graph construction scheme proposed in this application is geared towards the operation and maintenance of business systems, and the technical approach to its construction is as follows: Figure 8 As shown.

[0131] First, because the operation and maintenance of business systems involves many aspects and a large amount of related knowledge, staff need to organize and summarize the relevant knowledge in advance to form original domain knowledge content. Before constructing the domain knowledge graph, the domain of the knowledge graph needs to be determined according to the application direction. The schema layer design of the domain knowledge graph is mainly represented in the form of an ontology, mainly composed of concepts (nodes) and relations (connections between nodes). The purpose is to abstract specific domain knowledge into a knowledge representation framework. The ontology modeling in this application can be completed in Protégé, while the storage and fusion of the complete knowledge graph are carried out in a graph database, which involves the mapping from ontology to graph database. Specifically:

[0132] (1) Ontology modeling

[0133] Combination Figure 1 As shown, in the process of domain ontology modeling, determining the professional domain and scope of the domain ontology, sorting out and listing the important terms in the domain involved in the ontology, extracting concepts and classifying concept hierarchies, and defining the relationships between concepts are important steps. In this application, researchers can adopt a strategy combining deduction and induction. First, they can perform top-down deduction based on the divided domains; second, they can use bottom-up induction to sort out and summarize the actual domain knowledge; then, they can iteratively adjust the top-down and bottom-up methods to obtain the final domain ontology. The main tool used is Protege, in which concepts, instances, and relationships are imported, as well as constraints and attributes are defined, ultimately constructing the domain ontology (i.e., the domain ontology model mentioned above).

[0134] (2) Mapping rules

[0135] This application uses the graph characteristics of graph databases to describe the network structure of an ontology, and defines the mapping rules between the basic components of the ontology and the node storage elements of nodes in the graph database, as shown in Table 1.

[0136] Based on the mapping rules shown in Table 1, the OWL representation method of the ontology model (i.e., the domain ontology) can be embedded and stored in the graph database (domain graph data output through Neo4j) in this application, thereby completing the initial construction of the domain knowledge base (i.e., domain graph data).

[0137] The main content of the case knowledge graph is historical operation and maintenance cases. Therefore, it differs from the construction concept of the domain knowledge graph. It has lower requirements for the schema layer and higher requirements for the instance layer, requiring a knowledge graph-style description of an operation and maintenance event.

[0138] Since the components of the case knowledge graph are mostly "events", this application models each event as a graph structure "community", which requires a central node, analogous to the "primary key" in a relational database (due to the heterogeneity of data sources, the definition of the central node is different for each event, but the model structure is similar). The nodes that are directly related to the central node represent the relevant operation and maintenance information of the event, such as the fault phenomenon, the cause of the fault, and the final handling measures of the event.

[0139] The knowledge graph of cases mainly comes from two forms: tabular records and document records. Tabular records are mainly summaries of a series of operation and maintenance activities. They are characterized by being relatively structured and easy to import into graph databases, but the information they express is limited. On the other hand, document records are detailed records of each fault event from beginning to end. They are characterized by expressing extremely rich semantic information and can profoundly reflect the operation and maintenance experience of experts, but the data is unstructured.

[0140] For structured data recorded in tables, the features of each column can be used as relation names in a knowledge graph, and each cell can be imported into the graph database as a node of an event element. For unstructured data recorded in documents, preprocessing is required using Natural Language Processing (NLP) techniques. The main technical approaches are as follows: Figure 9 As shown, the original document is first converted into a text document (word2txt, i.e., the original document in word format is converted into a text document in txt format). Then, a natural language processing model based on the language representation model BERT (Bidirectional Encoder Representation from Transformers) is used to extract triples (including three subtasks: word serialization, entity recognition, and relation extraction). After manual verification, it is stored in a graph database (i.e., the subgraph data corresponding to each operation and maintenance event is stored in the graph database).

[0141] After obtaining the case knowledge graphs for each event (i.e., the subgraph data corresponding to each operation and maintenance event), the knowledge graphs are merged, and the same concepts are aligned to form the final case knowledge graph.

[0142] Finally, since the domain knowledge graph and case knowledge graph are constructed independently at this point, they need to be further integrated. Because the domain knowledge graph focuses more on the design of the schema layer, aiming to categorize all instances in the operation and maintenance records into a single class through top-down deduction, it is only necessary to summarize and classify all entities in the case knowledge graph from the bottom up, establishing "instance" relationships with the concepts in the domain knowledge graph, thus forming the entire system operation and maintenance domain knowledge base.

[0143] In summary, the technical solution proposed in this application has the following advantages:

[0144] First, this application employs a knowledge management approach. Compared to other knowledge representation methods, knowledge graphs store information in a graph structure, which is more intuitive, improves relation query efficiency, enhances semantic expression, and increases the efficiency of knowledge retrieval.

[0145] Secondly, this application employs a knowledge base construction method. By combining top-down deduction with bottom-up induction, the knowledge graph possesses both a rigorous and complete schema layer and a rich and diverse data layer, thereby minimizing information loss during the structured storage process of the knowledge system.

[0146] Finally, this application employs a knowledge storage tool. Compared to relational databases, graph databases offer more efficient relational queries and support semantic representation. Therefore, utilizing knowledge graphs for fault diagnosis allows the application of historical maintenance processes and expert experience to the maintenance workflow, making fault location and maintenance decisions more "intelligent."

[0147] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0148] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0149] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0150] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing an operations and maintenance knowledge graph, characterized in that, The method includes: Obtain domain knowledge content of the target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs; Based on the domain knowledge content, a domain graph data is constructed. The domain graph data contains multiple first nodes. Any two first nodes are directly connected or indirectly connected through other first nodes. The first node represents the domain knowledge content. Based on the operation and maintenance case content, case graph data is constructed. The case graph data consists of at least one sub-graph data. The sub-graph data are connected through a central node. The sub-graph data contains multiple second nodes. Any two second nodes are directly connected or indirectly connected through other second nodes. The second node represents the operation and maintenance case content to which it belongs. The domain graph data and the case graph data are fused to obtain the operation and maintenance knowledge graph corresponding to the target domain. The operation and maintenance case content includes table records and document records of the corresponding operation and maintenance events; Specifically, based on the operation and maintenance case content corresponding to the operation and maintenance event, subgraph data corresponding to the operation and maintenance event is constructed, including: Based on the table record content corresponding to the operation and maintenance event, construct the first graph data corresponding to the operation and maintenance event; Based on the document record content corresponding to the operation and maintenance event, construct the second graph data corresponding to the operation and maintenance event; Align the nodes in the first graph data and the nodes in the second graph data in terms of semantic description; The first graph data and the second graph data are merged to obtain the sub-graph data corresponding to the operation and maintenance event; The domain graph data is constructed based on the domain knowledge content, including: Based on a pre-set ontology model architecture, the domain knowledge content is processed to obtain a domain ontology model corresponding to the domain knowledge content. The domain ontology model contains multiple domain ontology. According to the conversion relationship in the preset mapping rules, the ontology components in the domain ontology are converted into node storage elements in the graph data nodes to obtain the first node based on the graph data, and the first node constitutes the domain graph data; wherein, the mapping rules include: the conversion relationship between the ontology components in the ontology and the node storage elements in the nodes; The step of fusing the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain includes: Align the nodes in the domain graph data and the nodes in the case graph data in terms of semantic description; Establish a parent-child relationship between at least one node in the case graph data and a node in the domain graph data, so that at least one node in the case graph data has a parent node in the domain graph data, to obtain the operation and maintenance knowledge graph corresponding to the target domain; Among them, the node storage elements of the two nodes that have established the parent-child relationship satisfy the node association condition.

2. The method according to claim 1, characterized in that, Each of the aforementioned operation and maintenance case contents corresponds to one operation and maintenance event; The step of constructing case graph data based on the content of the operation and maintenance case includes: Based on the operation and maintenance case content corresponding to the operation and maintenance event, construct subgraph data corresponding to the operation and maintenance event; each subgraph data corresponding to the operation and maintenance event has a root node, and the second node in each subgraph data is directly or indirectly connected to the root node; The root nodes of all the subgraph data are concatenated through the central node to obtain the case graph data.

3. The method according to claim 2, characterized in that, The operation and maintenance case content includes the table record content of the corresponding operation and maintenance events, and the table record content consists of the content of multiple cells; The step of constructing the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event includes: Using the cells in the table records as nodes of graph data, subgraph data corresponding to the operation and maintenance event is constructed based on the content of the cells.

4. The method according to claim 2, characterized in that, The operation and maintenance case content includes the document record content of the corresponding operation and maintenance event, and the document record content includes at least one record statement; The step of constructing the subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content corresponding to the operation and maintenance event includes: The record statements corresponding to the operation and maintenance events are extracted into triples to obtain multiple entities and the relationships between the entities; Based on the entities and the relationships between them, construct the subgraph data corresponding to the operation and maintenance events.

5. A device for constructing an operation and maintenance knowledge graph, characterized in that, The device includes: The content acquisition unit is used to acquire domain knowledge content of the target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs; A domain graph construction unit is used to construct domain graph data based on the domain knowledge content. The domain graph data includes multiple first nodes, and any first node has a connection relationship with at least one other first node. The first node represents the domain knowledge content. The case diagram construction unit is used to construct case diagram data based on the operation and maintenance case content. The case diagram data consists of at least one sub-diagram data, which are connected to each other through a central node. The sub-diagram data contains multiple second nodes, and any second node has a connection relationship with at least one other second node. The second node represents the operation and maintenance case content to which it belongs. The graph fusion unit is used to perform graph data fusion processing on the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain; The operation and maintenance case content includes table records and document records of the corresponding operation and maintenance events; Specifically, constructing subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content includes: constructing first graph data corresponding to the operation and maintenance event based on the table record content corresponding to the operation and maintenance event; constructing second graph data corresponding to the operation and maintenance event based on the document record content corresponding to the operation and maintenance event; aligning the nodes in the first graph data and the nodes in the second graph data in terms of semantic description; and fusing the first graph data and the second graph data to obtain the subgraph data corresponding to the operation and maintenance event. The process of constructing domain graph data based on the domain knowledge content includes: processing the domain knowledge content according to a pre-set ontology model architecture to obtain a domain ontology model corresponding to the domain knowledge content, wherein the domain ontology model contains multiple domain ontologs; and converting the ontology components in the domain ontologs into node storage elements in the graph data nodes according to the conversion relationship in the preset mapping rules to obtain a first node based on the graph data, wherein the first node constitutes the domain graph data; wherein the mapping rules include the conversion relationship between the ontology components in the ontologs and the node storage elements in the nodes. The step of fusing the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain includes: aligning the nodes in the domain graph data and the nodes in the case graph data in terms of semantic description; establishing a parent-child relationship between at least one node in the case graph data and a node in the domain graph data, such that at least one node in the case graph data has a parent node in the domain graph data, to obtain the operation and maintenance knowledge graph corresponding to the target domain; wherein the node storage elements of the two nodes with the established parent-child relationship satisfy the node association condition.

6. An electronic device, characterized in that, include: A memory used to store computer programs and the data generated by the execution of said computer programs; A processor is configured to execute the computer program to: obtain domain knowledge content of a target domain and at least one operation and maintenance case content of the target domain, wherein the target domain is the domain to which the operation and maintenance system belongs; Based on the domain knowledge content, a domain graph data is constructed, which contains multiple first nodes. Each first node is connected to at least one other first node, and each first node represents the domain knowledge content. Based on the operation and maintenance case content, case graph data is constructed, which consists of at least one subgraph data. The subgraph data are connected through a central node. Each subgraph data contains multiple second nodes. Each second node is connected to at least one other second node, and each second node represents the operation and maintenance case content to which it belongs. The domain graph data and the case graph data are fused to obtain the operation and maintenance knowledge graph corresponding to the target domain. The operation and maintenance case content includes table records and document records of the corresponding operation and maintenance events; Specifically, constructing subgraph data corresponding to the operation and maintenance event based on the operation and maintenance case content includes: constructing first graph data corresponding to the operation and maintenance event based on the table record content corresponding to the operation and maintenance event; constructing second graph data corresponding to the operation and maintenance event based on the document record content corresponding to the operation and maintenance event; aligning the nodes in the first graph data and the nodes in the second graph data in terms of semantic description; and fusing the first graph data and the second graph data to obtain the subgraph data corresponding to the operation and maintenance event. The process of constructing domain graph data based on the domain knowledge content includes: processing the domain knowledge content according to a pre-set ontology model architecture to obtain a domain ontology model corresponding to the domain knowledge content, wherein the domain ontology model contains multiple domain ontologs; and converting the ontology components in the domain ontologs into node storage elements in the graph data nodes according to the conversion relationship in the preset mapping rules to obtain a first node based on the graph data, wherein the first node constitutes the domain graph data; wherein the mapping rules include the conversion relationship between the ontology components in the ontologs and the node storage elements in the nodes. The step of fusing the domain graph data and the case graph data to obtain the operation and maintenance knowledge graph corresponding to the target domain includes: aligning the nodes in the domain graph data and the nodes in the case graph data in terms of semantic description; establishing a parent-child relationship between at least one node in the case graph data and a node in the domain graph data, such that at least one node in the case graph data has a parent node in the domain graph data, to obtain the operation and maintenance knowledge graph corresponding to the target domain; wherein the node storage elements of the two nodes with the established parent-child relationship satisfy the node association condition.

Images