A power grid ontology model construction method based on a graph model correlation recommendation algorithm

Abstract

The application discloses a power grid ontology model construction method based on a graph model association recommendation algorithm. The method can improve the construction efficiency of the power grid ontology model. The method comprises the following steps: step A: obtaining constructor information and modeling keywords of a model to be constructed; step B: calling multiple historical models from a historical model database; step C: identifying a concept matched with the modeling keywords in the called historical models as a keyword concept, configuring the constructor information as an attribute of the keyword concept of the historical model, and taking a concept corresponding to the attribute in the historical model as an associated concept; step D: taking multiple associated concepts and keyword concepts connected with each other as a group of association relationships, calculating the total number of association relationships of the historical model to obtain a total number of relationships; step E: taking the historical model with the total number of relationships reaching a preset degree in all the called historical models as a reference model; and step F: generating a power grid ontology model according to a selection operation of a user on the reference model.

Description

TECHNICAL FIELD

[0001] The present application mainly relates to the technical field of ontology model construction, and particularly relates to a power grid ontology model construction method based on a graph model association recommendation algorithm. BACKGROUND

[0002] The dynamic ontology modeling tool is realized by using big data analysis technology and knowledge graph technology, and is an open and independent application platform meeting the needs of data analysis, semantic fusion, business exploration and the like. The graph database is a storage engine specially used for storing and retrieving a huge relational information network, and can efficiently store data as nodes and the relationship between data as edges, and can add relevant attributes to the nodes or edges, and can realize efficient full-text retrieval and query of the nodes, edges and attributes. Figure 1 In the dynamic ontology modeling tool, the graph database mainly solves the storage of ontology concept objects, including storage of graph pattern information, semantic information, object attributes and object relationships. Users can realize high-performance retrieval and mining through graph retrieval languages such as Gremlin, TinkerPop, Cypher and nGQL, and the system can perform graph calculation and analysis through a distributed graph processing framework such as SparkGraphX.

[0003] Taking the construction of a power grid ontology model using the dynamic ontology modeling tool as an example, the power system management team involves procurement evaluation, operation and maintenance management, fault handling and the like when managing power equipment, and each management behavior needs to separately establish a power grid ontology model. However, the grass-roots management personnel need to spend a lot of effort to understand and sort out the required concepts and the relationship between the concepts in the process of constructing the model due to relatively insufficient information acquisition and relatively insufficient business experience, resulting in problems such as missing or incomplete expression of business-related information in the constructed ontology model. Therefore, the grass-roots management personnel often need to load a large number of instances to verify whether the model has problems, and the final effective model can be obtained only after repeated modification, which has a high human cost. SUMMARY

[0004] The technical problem to be solved by the present application is how to reduce the human cost in constructing a power grid ontology model.

[0005] In order to solve the above technical problem, the present application provides a power grid ontology model construction method based on a graph model association recommendation algorithm, comprising the following steps:

[0006] Step A: obtaining construction request information of a model to be constructed, the construction request information including current constructor information and modeling keywords;

[0007] Step B: retrieving a plurality of historical models from a pre-stored historical model database, which are matched with both the acquired current constructor information and the modeling keywords;

[0008] Step C: identifying the concepts in the retrieved historical models which are matched with the modeling keywords as keyword concepts, calling the keyword concepts, configuring the current constructor information as the attribute of the keyword concepts of each historical model, and respectively retrieving the concepts corresponding to the attribute from each historical model as associated concepts;

[0009] Step D: if there are a plurality of concepts connected with each other in all the associated concepts and the keyword concepts of a historical model, it is considered that there is a set of association relationships among the concepts, the total number of the association relationships of the historical model is calculated to obtain a relationship total number, and the greater the value of the relationship total number is, the higher the closeness between the historical model and the to-be-constructed model is;

[0010] Step E: calculating the relationship total number of each of the retrieved historical models respectively, and recommending the historical models whose relationship total number reaches a preset degree as reference models of the to-be-constructed model;

[0011] Step F: generating a power grid ontology model according to the selection operation of the user on the reference model.

[0012] Further, in the step B, the matched historical models refer to the historical models associated with the current constructor information and taking at least one of the modeling keywords as a concept.

[0013] Further, the constructor information includes a plurality of pre-defined label features, and the association with the current constructor information refers to that there are at least two label features in the constructor information of the historical model which are consistent with the label features of the current constructor information.

[0014] Further, the constructor information includes a plurality of pre-defined label features, and the concept and the constructor information both include a pre-defined category, the category of the keyword concept is identified, it is judged whether the label features belonging to the category in the constructor information of the to-be-constructed model and the historical model are matched respectively, and if the judgment result is matched, the value of the relationship total number associated with the keyword concept is enlarged by a preset increment.

[0015] Further, the category includes an entity, an event or a document.

[0016] Further, the constructor information includes one or more of the name, the unit, the job level, the work group and the name of the project participated by the constructor.

[0017] Further, in step D, the number of edges in a set of associated relationships is used as the degree value of the associated relationship, and the total number of associated relationships of the historical model is calculated, which means the total number of associated relationships whose degree values are within the preset threshold.

[0018] Further, if the total number of relationships of the two historical models calculated is the same, the historical model with more associated relationships whose degree values are outside the preset threshold has a larger value of the total number of relationships.

[0019] Further, in step D, the plurality of concepts are connected to each other, which means that these concepts are in a closed loop of a cycle.

[0020] Further, in step D, the plurality of concepts are connected to each other, which means that these concepts are in a closed loop and there are a starting point and an ending point of the closed loop in these concepts.

[0021] Further, in step D, the plurality of concepts connected to each other by the associated relationship include a keyword concept.

[0022] Further, in step D, the associated concept means a concept whose number of connected concepts in the historical model reaches a preset lower limit.

[0023] Beneficial effects of the invention: The above-mentioned method for constructing a power grid ontology model based on a graph model association recommendation algorithm retrieves multiple historical models from a pre-stored historical model database that are associated with the model to be constructed in terms of current builder information and modeling keywords; for each retrieved historical model, all keyword concepts corresponding to the modeling keywords are retrieved from its model concepts, and the current builder information is configured as an attribute of the keyword concept; then, the associated concepts corresponding to the attributes are retrieved. In this way, the builder of the current model is associated with the model content of the historical models through the associated concepts; and finally, the degree of association of these keyword concepts and associated concepts in the graph model is used as the basis for... To assess the affinity between historical models and the model to be constructed, specifically, based on a graph model association algorithm, any set of concepts directly or indirectly connected across all keyword concepts and related concepts is considered an association. The more numerous and complex the relationships between keyword concepts and related concepts related to the model to be constructed in the historical model, the larger the total number of relationships calculated for that historical model. This indicates a higher affinity between the historical model and the model to be constructed, making it a better reference for the model to be constructed. It is even possible to directly copy or slightly adjust the historical model to construct a new power grid ontology model. This power grid ontology model construction method calculates the total number of relationships between keyword concepts and related concepts related to the model to be constructed in the historical model based on a graph model association recommendation algorithm. It can directly recommend historical models with higher affinity to the model to be constructed as reference models. When constructing a power grid ontology model, users can reuse the concepts and conceptual relationships of the reference model, or even directly copy the reference model, reducing manpower consumption during the construction of the power grid ontology model and improving the construction efficiency. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the existing graph model data storage structure;

[0025] Figure 2 This is a flowchart of the power grid ontology model construction method based on graph model association recommendation algorithm provided by the present invention.

[0026] Figure 3 This is a schematic diagram of the model building function area interface in the dynamic ontology modeling tool provided by this invention.

[0027] Figure 4 This is a schematic diagram of a historical model 1 extracted based on a graph model association recommendation algorithm, which is the power grid ontology model construction method provided by this invention.

[0028] Figure 5 This is a schematic diagram of a historical model 2 extracted based on a graph model association recommendation algorithm, which is the power grid ontology model construction method provided by this invention.

[0029] Figure 6 is a schematic diagram of a history model three extracted by the power grid ontology model construction method provided by the application based on a graph model association recommendation algorithm.

[0030] Figure 7 is a schematic diagram of the current constructor information of the model to be constructed being configured to the modeling keywords of the history model in the form of label features in the power grid ontology model construction method provided by the application.

[0031] Figures 8-17 is a plurality of association relationship structures obtained by searching the history model three according to the power grid ontology model construction method provided by the application. Figure 6

[0032] Figure 18 is a schematic diagram of the constructor information relationship matching between the model to be constructed and the history model three in the power grid ontology model construction method based on the graph model association recommendation algorithm provided by the application. Figure 6

[0033] Figure 19 is a power grid ontology model constructed based on the history model three in the power grid ontology model construction method based on the graph model association recommendation algorithm provided by the application. Figure 6 DETAILED DESCRIPTION

[0034] The application will be further described in detail below in combination with specific embodiments.

[0035] The power grid ontology model construction method based on the graph model association recommendation algorithm of the present embodiment is shown in Figure 2 , and the specific steps of the above method will be described below by taking the process of constructing a power grid ontology model by a user on a dynamic ontology modeling tool as an example.

[0036] A. Obtain the construction request information of the model to be constructed, and the construction request information includes the current constructor information and the modeling keywords.

[0037] The user logs in to the dynamic ontology modeling tool and enters the model construction function area as shown in Figure 3 . The user submits a model construction request to the server of the dynamic ontology modeling tool, and the model to be constructed is “the transformer maintenance model of the regional power distribution network”.

[0038] The server reads the login user information as the current constructor information of the model to be constructed, and the constructor information of the present embodiment includes the name (Zhang), unit (power grid company M), department (equipment center), work group (equipment operation and maintenance group), job level (group leader), project (power grid equipment supervision), and task (transformer maintenance) of the user.

[0039] ​​​The server starts the power grid ontology model construction task after receiving the model construction request submitted by the user, extracts the modeling keywords: region, power transformation equipment and maintenance, and the specific keyword extraction algorithm is realized by the prior art, which is not described here.

[0040] B. retrieves a plurality of historical models from the pre-stored historical model database that match both the obtained current constructor information and the modeling keywords.

[0041] The historical model database of the embodiment records the constructor information of each historical model, which is pre-processed by semantic fusion normalization / standardization (realized by the prior art, which is not described here), and then stored in the server database in the form of data associated with the label features of the historical model. After reading the current constructor information of the to-be-constructed model, the server displays all the read constructor information to the user for the user to select to generate the label features of the constructor information of the to-be-constructed model. Specifically, the user can select multiple constructor information combinations to generate a label feature, such as selecting the combination of unit + department (power grid company M-equipment center) as a label feature; or directly selecting a single constructor information as a label feature, such as selecting a work group (equipment operation and maintenance group), a project (power grid equipment supervision) and a task (transformer maintenance) as three other label features.

[0042] The historical model of the embodiment matches the obtained current constructor information, for example: there are at least two label features in the constructor information of the historical model that are associated with the label features of the current constructor information, and the two label features are consistent or have a repetition rate higher than a preset value (such as 80%). The historical model of the embodiment matches the obtained modeling keywords, that is, the historical model has a concept consistent with or having a repetition rate higher than a preset value (such as 90%) with the modeling keywords. The modeling keywords are also pre-processed by semantic fusion normalization / standardization (realized by the prior art, which is not described here).

[0043] The server first reads the above four tag features of the current builder information, and performs a screening task of the tag features associated with the historical model in the self-database. The screening process does not need to call the model and read the model content, so it can quickly preliminarily screen the historical model associated with the user, i.e. the builder of the to-be-constructed model, from the historical model database. Then, the server releases the calling permission of all historical models associated with any two of the tag features in the historical model database, and then searches the three modeling keywords "region, power transformation equipment and maintenance" of the to-be-constructed model in all historical models with released permission. If there is a concept matching one of the three modeling keywords in the searched historical model, the historical model is marked as a candidate model of the to-be-constructed model and is called from the historical model database to a new candidate database. If the number of the marked candidate models of the to-be-constructed model exceeds the preset lower limit (such as 100), the search standard is improved for re-searching, such as marking the historical model as a candidate model only if there is a concept matching two or more modeling keywords in the historical model.

[0044] In the embodiment, the historical models called to the candidate database include a historical model one "power transformation equipment failure handling model" of Figure 4 , a historical model two "transformer procurement model" of Figure 5 , and a historical model three "power transformation equipment supervision model" of Figure 6 .

[0045] C. Identify the concepts matching the modeling keywords in the called historical models as keyword concepts, call the keyword concepts, configure the current builder information as the attribute of the keyword concepts of each historical model, and search the concepts corresponding to the attribute as associated concepts from each historical model.

[0046] Taking the historical model three "power transformation equipment supervision model" of Figure 6 as an example, the concepts of the historical model: region location, power transformation equipment and maintenance correspond to the three modeling keywords of the to-be-constructed model "power transformation equipment maintenance model of regional power distribution network": region, power transformation equipment and maintenance. See Figure 7The server configures the current constructor information of the model to be constructed as the attribute of the three modeling keywords in the form of label features, and based on the association algorithm of the graph model, the graph model retrieval tool is used to retrieve the concept corresponding to the attribute (label feature) associated with the keyword concept in the three historical models as the associated concept. The associated keyword concept means that it is directly connected with at least one associated keyword concept, and the attribute (label feature) corresponds to the concept that the attribute covers all label features or at least covers two label features. In this embodiment, the constructor information of the historical model is simulated to be configured as the attribute of the keyword concept of the historical model, and the constructor information originally only as an external additional feature of the model is configured as a keyword concept attribute in the historical model as a retrieval parameter in the model, so that the concept related to the constructor of the model to be constructed in the model can be accurately retrieved as the associated concept.

[0047] The server retrieves the associated concept corresponding to the four label feature attributes of the power grid company M-equipment center, equipment operation and maintenance group, power grid equipment supervision, and transformer maintenance in the historical model three shown in Figure 6 The server retrieves the associated concept corresponding to the four label feature attributes of the power grid company M-equipment center, equipment operation and maintenance group, power grid equipment supervision, and transformer maintenance in the historical model three shown in

[0048] D. If there are multiple concepts connected to each other in all associated concepts and keyword concepts of a historical model, it is considered that these concepts have a group of association relationships, the total number of relationships is calculated by calculating the total number of relationships of the historical model, and the greater the value of the total number of relationships, the higher the closeness of the historical model to the model to be constructed.

[0049] The server retrieves the associated concept corresponding to the four label feature attributes of the power grid company M-equipment center, equipment operation and maintenance group, power grid equipment supervision, and transformer maintenance in the historical model three shown in

[0050] Among them, the server specifically limits at least one keyword concept in these important concepts, and can also limit a specified keyword concept in these important concepts in the modeling purpose-specific modeling task. The non-modeling keyword concept whose connected concept number in the historical model reaches the preset lower limit (such as seven) can also be used as the supplementary keyword concept of the historical model to expand the model depth.

[0051] In addition, in the present embodiment, the number of all edges connecting adjacent two concepts (vertices) in the association relationship is taken as the degree value of the association relationship, such as Figure 1 The minimum number of vertices 2 and two vertices 3 between vertex 1 and vertex 42 on the right side can form a closed loop, and the connection between vertex 1 and vertex 42 is realized by at least six edges, so the minimum degree value of the association relationship is six, and the degree value of the association relationship obtained by detouring through other vertices exceeds six. Obviously, the association relationship will deviate from vertex 1 and vertex 42 more, so when calculating all the association relationships of the historical model, all the association relationships whose degree values are within the preset threshold (such as ≤6) are calculated. Further, if the total number of relationships of the two historical models calculated is the same, it is considered that the historical model with more association relationships with degree values outside the preset threshold has a larger value of the total number of relationships.

[0052] The server retrieves in the retrieved historical model by taking whether multiple concepts in all keyword concepts and associated concepts are connected with each other as the retrieval condition, that is, the server retrieves the historical model in which multiple important concepts are directly or indirectly connected together, such as in a closed loop in a cycle, or in a closed loop and the start and end points of the closed loop exist in the concepts, so it is considered that there is a set of association relationships between the important concepts in the historical model. The server calculates the number of all association relationships of the historical model to obtain the total number of relationships, and the larger the value of the total number of relationships is, the higher the closeness of the historical model to the to-be-constructed model is. The more and more complex the relationship between the keyword concepts and the associated concepts related to the to-be-constructed model in the historical model is, the larger the value of the total number of relationships obtained by calculating the number of all association relationships of the historical model is, so the closeness of the historical model to the to-be-constructed model is higher, which can provide a reference for the to-be-constructed model, and even the historical model can be directly copied or slightly adjusted to construct a new power grid ontology model.

[0053] In the present embodiment, Figure 6The illustrated historical model three (transformer equipment supervision model) is taken as an example to specifically describe the above step D. The server retrieves all the association relationships of the historical model with the number of covered associated concepts and keyword concepts as the first condition, and the degree value from small to large as the second condition, and the same association relationship is not repeated. The specific algorithm logic of the association relationship retrieval can adopt a path search algorithm such as an extended random tree, which is not described here. The keyword concept includes regional location, transformer equipment and maintenance, and the associated concept is point inspection event. Figure 6 The arrow of the connection (edge) between the two concepts, such as concept 1→concept 2, represents that there is a data mapping relationship mapped from concept 1 to concept 2, that is, the arrow direction in the figure represents that the next concept can be pointed to from this concept. The association relationship calculation of the historical model three (transformer equipment supervision model) is as follows.

[0054] (1) Four concepts: regional location, transformer equipment, maintenance, and point inspection event, have 1 association relationship, such as Figure 8 , the degree value is 7, which exceeds 6, and is not counted in the total number of relationships.

[0055] (2) Three concepts: regional location, transformer equipment, and maintenance, have 3 association relationships. There is 1 association relationship with a degree value of 6 (see Figure 9 ); replace “transformer equipment—occurrence→maintenance” in Figure 9 with “transformer equipment—

association

state

associated equipment

parameter

[0056] (3) Three concepts: regional location, transformer equipment, and point inspection event, do not have independent association relationships.

[0057] (4) Three concepts: regional location, maintenance, and point inspection event, do not have independent association relationships.

[0058] (5) Three concepts: transformer equipment, maintenance, and point inspection event, have 8 association relationships. There are 2 association relationships with a degree value of 3 (see Figure 11 ); there are 3 association relationships with a degree value of 4 (see Figure 12 ); and there are 3 association relationships with a degree value of 5 (see Figure 13 ).

[0059] (6) Two concepts: regional location only has 1 independent association relationship with transformer equipment, with a degree value of 3 (see Figure 14 ).

[0060] (7) Two concepts: transformer equipment and maintenance, have 11 association relationships. There are 3 association relationships with a degree value of 3 (see Figure 15 ), and there are 4 association relationships with a degree value of 4 (seeFigure 16 ), the Figure 16 “transformer equipment—occurrence→maintenance” in the two lower-layer association relations is replaced as in the above (2), obtaining four association relations with a degree value of 5 (as shown by the dashed lines in the lower layer). Figure 16

[0061] (8) Two concepts: transformer equipment and point inspection event, without independent association relations.

[0062] (9) Two concepts: maintenance and point inspection event, with two association relations (see Figure 17 ). One association relation has a degree value of 3, and one association relation has a degree value of 4.

[0063] The historical model three (transformer equipment supervision model) shown in the above Figure 6 has 22 valid association relations, that is, the value of the total number of relations of the historical model three is 22.

[0064] In the calculation of the total number of relations of the association relations of the above historical model three (transformer equipment supervision model), the association concepts and the supplemented keyword concepts are regarded as the same level concepts as the keyword concepts of the model actually needed to be constructed by the user, and the user is usually more concerned about the keyword concepts actually needed to be constructed, and therefore, in the graph model association recommendation algorithm of the embodiment, the total number of relations is fine-tuned by increasing the recommendation priority of the keyword concepts. Further, the closer the association relation between each historical model and the builder information of the user of the model to be constructed, the higher the recommendation priority of the graph model association recommendation algorithm should be, and the keyword concepts and the builder information of the historical model and the builder information of the model to be constructed are all pre-defined categories, so the embodiment associates the keyword concepts and the builder information by categories, thereby fine-tuning the total number of relations to recommend the historical model with higher intimacy to the user.

[0065] The categories of each builder information are pre-stored in the server, and the categories include entities, events or documents. In the model to be constructed, the name, unit and department of the builder information all belong to the category of entities, and the label features after the combination of multiple information are also entities. If multiple categories of builder information are combined to generate a label feature, the category of the label feature obtained after the combination needs to be manually selected by the user; the work group belongs to the entity; the project and the task both belong to the event. For each historical model, the categories of the concepts are also pre-defined. In the embodiment, a ring is used as the identifier of the entity category, a hexagon is used as the identifier of the event category, and a square is used as the identifier of the document category.

[0066] ​For all keyword concepts, the server identifies the category of each keyword concept and determines whether the tag features belonging to that category in the builder information of the model to be built and the historical model match. If the result is a match, the total number of relationships associated with that keyword concept is increased by a preset increment (e.g., 10%). All relationships starting from that keyword concept are multiplied by 110% to obtain the value of these relationships, and the total number of relationships in the historical model is recalculated. If the relationship of the keyword concept forms a closed loop, then the keyword concept is considered as a starting point for that relationship.

[0067] right Figure 6 The fine-tuning of the total number of relationships obtained from the above calculations in the historical model 3 (power equipment supervision model) is explained in detail below.

[0068] The key concepts in this embodiment include regional location (document), power equipment (entity), and maintenance (event), see [link to relevant documentation]. Figure 18 In the builder information of historical model three (transformer equipment supervision model), the tags that match the model to be built (or the current builder information of the model to be built) are the project's "equipment supervision" and the task's "transformer maintenance," both of which are events. Among all 22 relationships in this historical model, 14 historical models start with the concept "maintenance." Therefore, the total number of relationships in historical model three is: (22-14) + 14 * 110% = 23.4.

[0069] E. Perform the correlation calculation in step D above on all retrieved historical models (e.g., 30 models) to calculate the total number of relationships for each model.

[0070] like Figure 4 The historical model 1, "Transformer Equipment Fault Handling Model," has the keywords "transformer equipment" and "maintenance." The concept "fault" is an associated concept corresponding to the attribute "transformer maintenance" in the current builder information. Retrieving all relationships of historical model 1 using the three concepts "transformer equipment," "maintenance," and "fault," the total number of relationships is 11. "Transformer equipment" is an entity, and "maintenance" is an event. The event category "transformer maintenance" in the builder information of historical model 1 matches the builder information of the model to be built. Therefore, fine-tuning the total number of relationships using "maintenance" yields a final total of 11.4 relationships.

[0071] like Figure 5The shown historical model two "transformer procurement model" has the keyword concepts "transformer equipment", "regional location" and "maintenance", and the concept "procurement budget" of the historical model two connects another 8 concepts as the supplementary keyword concepts. The four concepts "transformer equipment", "regional location", "maintenance" and "procurement budget" are used to search all the association relations of the historical model two, and the value of the total number of relations calculated from the total number of relations of the historical model two is 7. The category of "transformer equipment" is entity, the category of "regional location" is document, and the category of "maintenance" is event. The "unit-department" in the constructor information of the historical model two belongs to the entity category and matches the constructor information of the model to be constructed, so the fine-tuning of the total number of relations with "transformer equipment" obtains the final total number of relations as 7.7.

[0072] The server recommends the historical model whose calculated total number of relations reaches a preset degree (such as more than 10 association relations) as the reference model of the model to be constructed, and if there are too many reference models meeting the requirements (such as more than 5), the server displays the first 5 reference models in order on the canvas in the order of the total number of relations from large to small. In this embodiment, the reference models include the historical model one and the historical model three, and the recommendation priority of the historical model three is higher than that of the historical model one.

[0073] F. Generating the power grid ontology model according to the selection operation of the user on the reference model. The server provides all the reference models to the user in a side-by-side display manner, and the user directly selects the historical model three by a point selection operation to reuse the historical model, and the server generates a new power grid ontology model "transformer equipment maintenance model of regional distribution network" on the canvas. The user manually adjusts the reused historical model on the canvas, such as deleting concepts, modifying concepts and the relations between concepts, and thus obtains a new power grid ontology model with the historical model as the reference.

[0074] The server has a preset model automatic arrangement function. For example, the server displays the historical model on the canvas Figure 6 , and the user selects the model automatically arranged by the server. The server searches the concepts and the relations between adjacent concepts in the model on the current canvas that do not exist in all the association relations of the historical model, marks these concepts / relations and reminds the user in a highlighted manner, and the user can choose to delete or retain these concepts / relations. The automatically arranged "transformer equipment maintenance model of regional distribution network" is shown in Figure 19 .

[0075] The power grid ontology model construction method based on the graph model association recommendation algorithm retrieves a plurality of historical models associated with the model to be constructed in terms of current constructor information and modeling keywords from a pre-stored historical model database; retrieves all keyword concepts corresponding to the modeling keywords from the model concepts of each historical model, and configures the current constructor information as an attribute of the keyword concepts, and then retrieves the associated concepts corresponding to the attribute, so that the constructor of the current model is associated with the model content of the historical model through the associated concepts; and the degree of association of these keyword concepts and associated concepts in the graph model is used as an index to judge the closeness of each historical model to the model to be constructed. The power grid ontology model construction method calculates the total number of relationships between the relevant keyword concepts and associated concepts in the historical model and the model to be constructed based on the graph model association recommendation algorithm, and can directly recommend the historical model with higher closeness to the model to be constructed as a reference model. The historical model with strong association with the model constructor and the modeling keywords (modeling purpose) is automatically analyzed and recommended to the user using the multi-dimensional association relationship between the concepts. When constructing the power grid ontology model, the user can use the concepts and concept relationships of the reference model, or even directly copy the reference model, reducing the human cost of constructing the power grid ontology model and improving the construction efficiency of the power grid ontology model.

[0076] The above is only an embodiment of the application, and does not limit the scope of patent protection. Those skilled in the art can make non-essential changes or substitutions based on the application, and still fall within the scope of patent protection.

Claims

1. A method for constructing a power grid ontology model based on a graph model-based association recommendation algorithm, characterized by: Includes the following steps: Step A: Obtain the build request information for the model to be built. The build request information includes the current builder information and modeling keywords; Step B: Retrieve multiple historical models from the pre-stored historical model database that match the current builder information and modeling keywords obtained; Step C: Identify concepts in the retrieved historical models that match the modeling keywords as keyword concepts. Concepts that match the modeling keywords refer to concepts that are consistent with the modeling keywords or have a repetition rate higher than a preset value. Call these keyword concepts, configure the current builder information as the attributes of the keyword concepts in each historical model, and retrieve the concepts corresponding to the attributes from each historical model as associated concepts. Concepts corresponding to the attributes refer to: based on the graph model association algorithm, the graph model retrieval tool retrieves the concepts in the historical models that are associated with these keyword concepts and correspond to the attributes through attribute retrieval. Step D: If multiple concepts are connected to each other in all the related concepts and keyword concepts of a historical model, then these concepts are considered to have a set of relationships. Calculate the total number of relationships in the historical model. The larger the total number of relationships, the higher the closeness between the historical model and the model to be constructed. Step E: For all retrieved historical models, calculate the total number of relations for each model, and recommend historical models whose total number of relations reaches a preset level as reference models for the model to be built; Step F: Generate the power grid ontology model based on the user's selection of the reference model.

2. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 1, characterized in that: In step B, the matching historical model refers to a historical model associated with the current builder information that uses at least one of the modeling keywords as a concept.

3. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 2, characterized in that: The builder information includes multiple predefined label features, and the association with the current builder information means that at least two label features in the builder information of the historical model are consistent with the label features of the current builder information.

4. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 1, characterized in that: The builder information includes multiple predefined label features. Both the concept and the builder information include predefined categories. The category of the keyword concept is identified. It is determined whether the label features belonging to the category in the builder information of the model to be built and the historical model match. If the determination result is a match, the total number of relationships associated with the keyword concept is increased by a preset increment.

5. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 4, characterized in that: The categories include entities, events, or documents.

6. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 1 or 4, characterized in that: The builder information includes one or more of the following: the builder's name, organization, job title, work group, and the name of the project they participated in.

7. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 1, characterized in that: In step D, the number of edges in a set of associations is used as the degree value of the association. Calculating all associations in the historical model refers to all associations whose degree values ​​are within a preset threshold.

8. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 7, characterized in that: If the total number of relationships in the two historical models is the same, the historical model with more relationships whose degree values ​​are outside the preset threshold will have a larger total number of relationships.

9. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 1, characterized in that: In step D, the interconnection of the multiple concepts means that these concepts are all within a closed loop.

10. The method for constructing a power grid ontology model based on a graph model association recommendation algorithm as described in claim 1, characterized in that: In step D, the interconnection of the multiple concepts means that these concepts are all within a closed loop and that the starting point and ending point of the closed loop exist within these concepts.

Images