Power data semantic association retrieval method based on field-level relationship evolution

Abstract

The application discloses a power data semantic association retrieval method based on field-level relationship evolution, relates to the technical field of power data intelligent processing, and comprises the following steps: collecting historical attributes, business descriptions, calling positions, frequencies and associated objects of power data fields, counting business direction change times and reference object change quantities, constructing field evolution tracks, and extracting change characteristics of multiple business directions corresponding to the same field to generate the field evolution tracks. Based on the field-level relationship evolution, the application constructs a dynamic semantic association mechanism, establishes independent association paths for different business semantics, and completes field association content reorganization and data content extraction according to business directions, so that the semantic association relationship can be continuously updated with business evolution, and the accuracy of power data semantic association retrieval, the business differentiation capability, and the data association credibility of intelligent question answering, knowledge reasoning and natural language query conversion are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent power data processing technology, specifically to a semantic association retrieval method for power data based on field-level relationship evolution. Background Technology

[0002] Semantic association retrieval of power data based on field-level relationship evolution refers to using fields in power data assets as the smallest semantic unit. It constructs a continuously updated field-level relationship network around multi-dimensional connections such as primary and foreign key relationships, business relationships, computational dependencies, lineage relationships, synonym relationships, and usage popularity relationships between fields. This network is then combined with business function paths, natural language semantics, multimodal information understanding, and relationship graph reasoning to achieve accurate positioning and intelligent retrieval of power data. Compared to traditional data retrieval methods that rely on table names, field names, or keyword matching, this method pays more attention to the dynamic changes in field semantic relationships caused by business evolution, system iteration, and data asset updates. By continuously maintaining the field relationship network, new fields, modified fields, and changes in business definitions are synchronously mapped to the relationship graph, maintaining the consistency of semantic relationships between fields. This technology uses the data asset catalog automatically generated in Task 1 as a foundation to further construct a field-level power data relationship graph. It integrates business function paths, interface screenshots, natural language descriptions, and graph inference results. It uses graph traversal, vector retrieval, and keyword retrieval to complete multi-path retrieval in collaboration. Then, it completes sorting optimization based on the continuously evolving relationships between fields. This enables the system to accurately establish the mapping relationship between business semantics and technical data, and achieve rapid data discovery and precise positioning across systems, data tables, and business domains. It provides unified semantic support and dynamic knowledge foundation for subsequent knowledge-enhanced natural language query conversion, reliable structured query generation, and intelligent question answering.

[0003] Existing technologies have the following shortcomings: Current power data semantic association retrieval systems typically build fixed data associations based on initial field definition information. As power businesses continue to expand and business scenarios change, the same field may carry different business entities or generate new business semantics. However, existing systems lack the ability to identify the semantic evolution process of fields and still use the initially established single semantic association path for retrieval. This results in the same query conditions corresponding to the same retrieval path under different business scenarios, failing to form differentiated data associations that match the current business semantics. Consequently, some real business data cannot be effectively discovered, reducing the accuracy of data retrieval and the credibility of subsequent intelligent analysis applications.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a semantic association retrieval method for power data based on field-level relationship evolution, so as to solve the problems in the background art mentioned above.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a semantic association retrieval method for power data based on field-level relationship evolution, comprising the following steps:

[0007] S001, collect historical attributes, business descriptions, call locations, reference frequencies, and associated objects of power data fields, count the number of business pointer changes and the number of referenced objects changes, construct field evolution trajectories, and extract the change characteristics of multiple business pointers corresponding to the same field to generate field evolution trajectories;

[0008] S002, locate the change position of each business pointer in the evolution trajectory of the positioning field, perform semantic decomposition on adjacent business pointers, generate a unique semantic identifier according to the difference of business pointers, configure a unique identification mark for each business pointer, and establish a semantic identifier correspondence;

[0009] S003, Read the semantic identifier correspondence, establish an independent association entry for each unique semantic identifier, generate corresponding association paths according to the number of business scope changes and the number of referenced objects changes, map each business pointer to the corresponding association entry, and construct association path information;

[0010] S004, Receive power data semantic association retrieval request, use association path information to match the fields in the query conditions to the corresponding unique identification mark, select the corresponding association path according to the degree of consistency of business direction, and determine the business association path;

[0011] S005: Based on the business association path, reorganize the field association content, extract the corresponding data content according to the business direction, and output the semantic association retrieval results of power data corresponding to the business association path.

[0012] Preferably, in step S001, when collecting the historical attributes, business descriptions, call locations, reference frequencies, and associated objects of power data fields, the historical records of the same field formed in different business stages are continuously organized according to the order of business occurrence. The historical attributes include field name, field type, data object to which the field belongs, field business description, and maintenance time. The number of business pointer changes is counted according to the changes in business descriptions, and the number of referenced objects is counted according to the changes in associated objects. The field evolution trajectory is constructed according to the continuous correspondence of the historical records. The field evolution trajectory is used to record the development process of the corresponding business pointers of the same field in multiple business stages and the continuous association between each business stage.

[0013] Preferably, in step S002, when locating the change positions of each business direction in the field evolution trajectory, adjacent business stages in the field evolution trajectory are grouped into business direction change units, and the change positions of business directions are determined according to the business description, call position, reference frequency and the corresponding relationship of associated objects; semantic decomposition is performed for each change position of business direction, the business semantics corresponding to the same field in different business stages are expressed independently, and unique semantic identifiers are generated according to the differences in business directions, so that different business directions correspond to different unique semantic identifiers.

[0014] Preferably, in step S002, when configuring a unique identification mark for each business pointer, the unique identification mark is respectively mapped to a unique semantic identifier and a business pointer, and the fixed correspondence between the unique identification mark and the business description, call location, reference frequency and associated object is maintained; a semantic identifier correspondence is established based on all unique identification marks, so that when the same field corresponds to multiple business pointers, each business pointer corresponds to an independent unique semantic identifier and a unique identification mark, and the developmental connection of each business pointer originating from the same field evolution trajectory is preserved.

[0015] Preferably, in step S003, when establishing an independent association entry for each unique semantic identifier, each independent association entry corresponds to a unique semantic identifier and saves the corresponding business direction, unique identification mark, number of business scope changes, and number of referenced object changes; corresponding association paths are generated based on the number of business scope changes and the number of referenced object changes, so that different unique semantic identifiers have independent association entry and independent association paths, and maintain the business development relationship originating from the same field evolution trajectory.

[0016] Preferably, in step S003, when mapping each business pointer to its corresponding associated entry point, a unique semantic identifier is used as the mapping basis, so that each business pointer corresponds to only one independent associated entry point, and the business description, call location, associated object and business scope change information are organized along the corresponding associated path to construct the associated path information; the associated path information includes a unique semantic identifier, a unique identification mark, an independent associated entry point, an associated path and the correspondence between business pointers.

[0017] Preferably, in step S004, after receiving the power data semantic association retrieval request, the fields pointed to in the query conditions are extracted, and a correspondence is established with the unique identification mark in the association path information based on the business description, call source, business scenario and associated object corresponding to the field pointed to, so that the fields in the query conditions point to the corresponding unique business semantics and form a corresponding candidate association path set.

[0018] Preferably, in step S004, when selecting the corresponding associated path according to the degree of consistency of business direction, the business correspondence between the field direction in the query conditions and the corresponding business direction of each candidate associated path is established respectively, and the degree of consistency of business direction is determined according to the corresponding status of business description, business scope, call source and associated object, and the corresponding business associated path is selected from the candidate associated paths.

[0019] Preferably, in step S005, when reorganizing the field association content based on the business association path, the historical attributes, business descriptions, call locations, associated objects, and business relationships between fields are reorganized according to the business directions saved in the business association path, so that the field association content is continuously organized along the business association path, different business association paths correspond to different field association content, and the business development relationship originating from the same field evolution trajectory is maintained.

[0020] Preferably, when extracting corresponding data content according to business direction, the business records, business objects and related data content corresponding to the fields are extracted based on the business direction corresponding to the business association path, and the correspondence between the data content and the field association content is maintained; based on the extracted data content, the power data semantic association retrieval results corresponding to the business association path are output, so that different business association paths correspond to different power data semantic association retrieval results.

[0021] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0022] This application uses fields as the basic object of semantic evolution, continuously collecting evolutionary information such as historical field attributes, business descriptions, call locations, reference frequencies, and associated objects. It constructs field evolution trajectories, further pinpoints changes in business focus, completes semantic decomposition and establishes unique semantic identifiers. Subsequently, it establishes independent association entry points and corresponding association paths for different business semantics, completing the reorganization of association relationships based on the continuous evolution of field business semantics. When the same field gradually evolves into multiple business meanings, each business meaning has an independent semantic expression and association path. The retrieval process can enter the corresponding business association path based on different business focuses, no longer relying on the initially established single semantic association relationship. Therefore, it can maintain the synchronous evolution of field business semantics and association relationships, improve the adaptability of power data semantic association retrieval to business changes, and enhance the accuracy of semantic association results under different business scenarios.

[0023] This application reorganizes field association content around the established business association paths, extracts corresponding data content according to business directions, and finally outputs power data semantic association retrieval results corresponding to the business association paths, maintaining a continuous correspondence between the field business semantic positioning process and the data association process. When different business scenarios input the same field query conditions, the retrieval request can select the corresponding business association path based on the business direction and output data content matching the business semantics. Different business needs correspond to different retrieval results, avoiding semantic confusion caused by different business scenarios sharing the same field association relationship for a long time. This provides a consistent data association foundation for intelligent question answering, knowledge reasoning, and natural language query conversion, while enhancing the business differentiation capability of power data semantic association retrieval results and the credibility of intelligent data application processes. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0025] Figure 1 This is a flowchart of the semantic association retrieval method for power data based on field-level relationship evolution according to the present invention. Detailed Implementation

[0026] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0027] This invention provides, for example Figure 1 The power data semantic association retrieval method based on field-level relationship evolution, as shown, includes the following steps:

[0028] S001, collect historical attributes, business descriptions, call locations, reference frequencies, and associated objects of power data fields, count the number of business pointer changes and the number of referenced objects changes, construct field evolution trajectories, and extract the change characteristics of multiple business pointers corresponding to the same field to generate field evolution trajectories;

[0029] The construction of field evolution trajectories reflects the semantic changes of power data fields during the continuous development of business operations. It records not only changes in the field's own attributes but also the continuous associations formed over time by the evolution of the field's business implications. This provides a unified data foundation for subsequently extracting the change features of multiple business implications corresponding to the same field. Therefore, the specific steps for generating field evolution trajectories are as follows: For the historical attributes, business descriptions, call locations, reference frequencies, and associated objects of the collected power data fields, the number of changes in business implications and the number of changes in referenced objects are statistically analyzed.

[0030] A continuous historical information collection process is established for each field in the power data, using the field's historical attributes, business descriptions, call locations, reference frequencies, and associated objects as historical records for the same field at different business stages. Specifically, field historical attributes include information reflecting the field's historical status, such as field name, field type, data object to which the field belongs, field business description, field business definition, and field maintenance time; business descriptions include the definition of the field's purpose at different business stages, business explanations, and the corresponding business description when the field participates in business processing; call locations include the actual location of the field in different business pages, business processes, business interfaces, business queries, and data processing stages; reference frequency records the cumulative number of times the field is referenced by different business activities within different business cycles; and associated objects include the data objects, business objects, calculation objects, and associated field sets corresponding to the field at each business stage.

[0031] All of the above information is continuously recorded according to the time of business occurrence, and a continuous correspondence is maintained between all historical records of the same field. Each new historical record is linked with the existing historical records, so that the same field forms a continuous change state in different business periods, providing a complete data source for subsequent identification of changes in business direction.

[0032] By utilizing continuously recorded historical field information, the business status of the same field at different historical stages is characterized by its business description, call location, reference frequency, and associated objects. A correspondence between each business stage is established according to the order of appearance in the historical records. In this process, the business pointer represents the actual business entity, business object, or business application direction corresponding to a field at a certain business stage. Changes in the business pointer are considered to occur when the business description changes, the call location corresponds to a new business scope, the associated object generates new business relationships, or the reference frequency continues to concentrate on new business activities. The location where the corresponding business pointer appears is recorded accordingly.

[0033] At the same time, the number of business direction changes is counted for each change, and the number of referenced objects is counted for the set of associated objects corresponding to each business stage. This allows the number of business direction changes to reflect the development process of the field's business purpose, and the number of referenced objects to reflect the development state of the field's association scope. These two types of change information jointly describe the continuous expansion of the business semantics of the same field, providing a complete basis for the field's evolution trajectory. The field evolution trajectory represents the continuous change process of the business semantics of the same field in the continuous business development process, formed by the sequential connection of multiple historical business stages. It not only includes the order in which the business direction changes, but also saves the historical state, association state, and change state corresponding to each business direction, thus fully reflecting the entire process of the field's business semantics evolving from the initial state to the current state.

[0034] Using the number of business pointer changes, the number of referenced objects changes, and the continuous change status in the field evolution trajectory as unified description content, the field evolution trajectory is continuously organized. The business description, call position, reference frequency, and associated objects corresponding to each business stage are configured to the corresponding business pointer position, so that each business stage in the field evolution trajectory has an independent business status description and maintains the continuous association between the business stages before and after.

[0035] When the same field goes through multiple business stages, the business stages are arranged in chronological order. The number of changes in business pointers corresponds to the number of times the business semantics have developed, and the number of changes in referenced objects corresponds to the development of the business association scope. Together, they reflect the development trend of the field's business semantics. At the same time, the field evolution trajectory retains the historical attributes, business descriptions, call locations, reference frequencies, and correspondences between associated objects for each business stage. This ensures that any business stage can fully reflect the business state of the field at that time, and that different business stages together constitute a complete trajectory of the continuous development of the field's business semantics. This avoids the mixing of multiple business stages and provides a continuous state expression basis for subsequent identification of multiple business pointers corresponding to the same field.

[0036] For the field evolution trajectory that has been continuously organized, the development relationship between each business direction is analyzed item by item in all business stages of the trajectory. The business description, call position, reference frequency and associated objects corresponding to different business stages are expressed as a whole. Multiple business directions corresponding to the same field in different historical stages are identified, and the development relationship between each business direction is recorded.

[0037] When multiple business stages correspond to different business purposes, the field evolution trajectory will fully retain the continuous change relationship corresponding to each business direction. At the same time, it will extract the change features of the same field corresponding to multiple business directions. These change features include business description change features, call location change features, reference frequency change features, associated object change features, and business direction change features. Each type of change feature corresponds to the development status of the field in different business stages and together constitutes a complete description of the field's business semantic evolution. Finally, the field evolution trajectory is generated, providing a continuous, complete, and unified business semantic evolution foundation for subsequent location of business direction change positions, semantic decomposition, and establishment of unique semantic identifiers.

[0038] S002, locate the change position of each business pointer in the evolution trajectory of the positioning field, perform semantic decomposition on adjacent business pointers, generate a unique semantic identifier according to the difference of business pointers, configure a unique identification mark for each business pointer, and establish a semantic identifier correspondence;

[0039] Semantic decomposition is used to identify different business meanings of the same field during continuous business evolution, and to establish independent semantic expressions for multiple business orientations to avoid multiple business meanings sharing the same semantic description for a long time. Based on the field evolution trajectory generated above, the position of the change in business orientation of the field in each business stage is located, and a unique semantic identifier is generated according to the difference in business orientation. A unique identification mark is configured for each business orientation, and finally, the semantic identifier correspondence is established. The specific steps are as follows:

[0040] Read all business stage records in the field evolution trajectory, and locate the change position of each business pointer item by item according to the business development order saved in the field evolution trajectory. Each change position of the business pointer corresponds to a historical node in which the business semantics of the field have changed. Therefore, record the business pointer before the change, the business pointer after the change, the corresponding business description, the call position, the reference frequency, and the associated object, and maintain a one-to-one correspondence between the contents.

[0041] When multiple business stages exist in the field evolution trajectory, these stages are arranged chronologically, with adjacent stages forming a business orientation change unit. Each business orientation change unit retains the field's business state across two consecutive business stages, describing the evolution of business semantics from one stage to the next. Simultaneously, a set of business orientation change records is established for all business orientation change locations, ensuring that each business orientation change in the field evolution trajectory corresponds to a unique change location number and maintaining a continuous association between the field evolution trajectory, business orientation change locations, and business states. For example, Table 1 provides an example of business orientation change location records.

[0042] Table 1: Example of Records of Changes in Business Location

[0043] P01 Marketing customers Integrated Energy Customers Integrated Energy Business Page Integrated Energy Orders P02 Integrated Energy Customers Charging customers Charging service page Charging orders P03 Electricity statistics Comprehensive energy consumption statistics Energy consumption analysis page Energy consumption analysis object

[0044] Each record in Table 1 corresponds to a business-oriented change point in the field's evolution trajectory, thus ensuring that subsequent semantic splitting is based on clear business change points, rather than directly performing unified processing on the field as a whole.

[0045] For each change in business pointer location, the differences in business meaning between adjacent business pointers are analyzed. The business pointer before and after the change are treated as two independent business semantics, and semantic decomposition is performed based on the business description, call location, reference frequency, and the business content jointly expressed by related objects. Semantic decomposition does not change the field itself, but rather independently expresses the business meaning of the same field at different business stages, so that one field corresponds to multiple independent business semantic units. After each semantic decomposition, a unique semantic identifier is generated for each decomposed business semantic. This unique semantic identifier corresponds to one business pointer, maintaining a state where different business semantics of the same field do not overlap or mix.

[0046] For example, if the same field represents a marketing customer in the marketing business stage and a comprehensive energy customer in the integrated energy business stage, then two unique semantic identifiers will be generated respectively. Each unique semantic identifier corresponds to an independent business semantic and will no longer share the same semantic description. Table 2 provides an example of unique semantic identifier generation.

[0047] Table 2: Examples of Unique Semantic Identifier Generation

[0048] customer_id Marketing customers S001 customer_id Integrated Energy Customers S002 customer_id Charging customers S003 power_type Electricity statistics S004 power_type Comprehensive energy consumption statistics S005

[0049] In Table 2, the same field can correspond to multiple unique semantic identifiers, and each unique semantic identifier corresponds to only one business pointer, thus completing the independent expression of the field's business semantics.

[0050] For each generated unique semantic identifier, a unique identification tag is configured for each business pointer, maintaining a fixed correspondence between the unique identification tag and the unique semantic identifier. The unique identification tag describes the unique identity of the business semantic in subsequent association processes. Its content not only records the unique semantic identifier but also retains the corresponding business description, call location, reference frequency, and the corresponding state between associated objects, ensuring that each unique identification tag can fully express the development background of the corresponding business semantic.

[0051] As the field's evolution trajectory continues to add new business stages, the newly generated business pointers continue to generate new unique semantic identifiers and configure new unique identification tags, while the original unique identification tags remain unchanged. This ensures that different business semantics of the same field are independently stored for a long time, without being overwritten or replaced by subsequent business expansions. Each unique identification tag remains independent of the others, but still retains its historical connection from the same field's evolution trajectory, allowing the development process of the field's business semantics to continue completely, while ensuring that each business semantic has a unique identity expression.

[0052] A semantic identifier correspondence is established using all unique identifiers. The corresponding content is continuously organized according to the correspondence status between fields, business directions, unique semantic identifiers, and unique identifiers, generating a complete semantic identifier correspondence. This semantic identifier correspondence represents the complete mapping relationship between different business directions and unique semantic identifiers for the same field during continuous business evolution. It not only preserves the one-to-one correspondence between fields and multiple business semantics but also preserves the development relationship of each business semantic from different business stages in the field's evolution trajectory. Therefore, it can completely express the development process of the same field gradually expanding from a single business semantic to multiple business semantics.

[0053] When new business directions emerge, only new unique semantic identifiers and unique identification tags are added to the semantic identifier correspondence, while existing correspondences remain stable. This constructs a continuously expanding, uniformly expressed, and mutually independent semantic identifier correspondence, providing a complete semantic mapping foundation for establishing independent associated entry points and generating corresponding associated paths. Table 3 provides an example of the semantic identifier correspondence.

[0054] Table 3: Examples of Semantic Identifier Correspondences

[0055] customer_id Marketing customers S001 M001 customer_id Integrated Energy Customers S002 M002 customer_id Charging customers S003 M003 power_type Electricity statistics S004 M004 power_type Comprehensive energy consumption statistics S005 M005

[0056] The semantic identifier correspondence shown in Table 3 fully records the correspondence between fields, business pointers, unique semantic identifiers, and unique identification tags, providing a unified and continuous semantic foundation for establishing differentiated association paths for different business scenarios in the future.

[0057] S003, Read the semantic identifier correspondence, establish an independent association entry for each unique semantic identifier, generate corresponding association paths according to the number of business scope changes and the number of referenced objects changes, map each business pointer to the corresponding association entry, and construct association path information;

[0058] The establishment of association paths is used to organize the unique semantic identifiers that have undergone semantic decomposition into independent business association expressions, so that different business pointers have their own corresponding association entry points, and to maintain an independent and continuous association state between business semantics. Based on the semantic identifier correspondence, an independent association entry point is established for each unique semantic identifier. Corresponding association paths are generated according to the number of times the business scope changes and the number of changes in the referenced objects. Each business pointer is mapped to its corresponding association entry point, and finally, the association path information is constructed. The specific steps are as follows:

[0059] Read all corresponding records in the semantic identifier mapping relationship, using the unique semantic identifier as the unique business semantic expression unit, and establish an independent association entry for each unique semantic identifier. The independent association entry represents the starting position corresponding to the unique semantic identifier when participating in business association. It maintains a fixed correspondence between the unique semantic identifier, business pointer, and unique identification mark, and records the initial state of the number of business scope changes and the number of referenced objects, ensuring that different business pointers correspond to different independent association entries, avoiding multiple business pointers sharing the same association starting position. When there are multiple unique semantic identifiers in the semantic identifier mapping relationship, each independent association entry is stored independently and a continuous organizational relationship is established according to the field source, providing a unified data foundation for subsequent association path generation. To quantify the business evolution state corresponding to each independent association entry, an association entry activity coefficient is introduced, calculated as follows:

[0060]

[0061] In the formula, This represents the activity coefficient of the associated entry point, used to indicate the basic degree of association when the corresponding unique semantic identifier participates in business association; This indicates the number of times the business scope has changed, that is, the cumulative number of times the field's business coverage has changed during the continuous evolution of the business; This indicates the number of changes to the referenced objects, that is, the cumulative number of changes to the objects associated with the field; the denominator contains... This represents the distance between the number of changes in the business scope and the number of changes in the referenced objects, reflecting the degree of coordination between the two types of changes. When the number of changes in the business scope and the number of changes in the referenced objects maintain a continuous correspondence, the activity coefficient of the associated entry point continues to increase, providing a unified quantitative basis for the generation of subsequent associated paths.

[0062] Based on the activity coefficient of each independent associated entry point, the development relationship between the number of changes in business scope and the number of changes in referenced objects is organized, and corresponding association paths are generated. An association path represents the development process of a unique semantic identifier starting from an independent associated entry point and gradually connecting business directions along the direction of continuous business evolution. It not only records the development order between business directions but also saves the business association state formed by the combined effect of the number of changes in business scope and the number of changes in referenced objects. Therefore, each association path corresponds to a unique semantic identifier, and different unique semantic identifiers form different association paths. To describe the development degree of the association path, an association path extension value is introduced, calculated as follows:

[0063]

[0064] In the formula, This represents the extension value of the associated path, used to describe the extent to which the associated path continues to expand along the direction of business evolution; Indicates the number of times the business scope has changed; This indicates the number of changes in the referenced object; the constant "1" ensures that the initial business state can also generate associated paths. This formula uses the activity coefficient of the associated entry point as the basis for path development, allowing both the number of changes in the business scope and the number of changes in the referenced object to participate in the organization of associated paths. As both types of changes continue to increase, the associated path extension value increases synchronously, which can more completely reflect the development state of the continuous expansion of business semantics and provide quantitative support for generating corresponding associated paths for different business directions.

[0065] Based on the association path extension value, each business pointer is mapped to its corresponding independent association entry point, ensuring that each business pointer enters an association path consistent with its unique semantic identifier, and maintaining a non-overlapping and non-mixing association relationship between business pointers. During the mapping process, multiple business pointers generated by the same field correspond to different independent association entry points. Each business pointer sequentially organizes its business description, call location, referenced object, and business development status along its corresponding association path, maintaining a continuous expression of the association path. Furthermore, to further describe the degree of matching between business pointers and association paths, a path association coverage coefficient is introduced, calculated as follows:

[0066]

[0067] In the formula, The path association coverage coefficient is used to indicate the degree of association coverage after a business pointer is mapped to the corresponding associated path. This indicates the number of services that the current associated path points to; Indicates the number of times the business scope has changed; This indicates the number of changes in the referenced objects. The path association coverage coefficient is based on the association path extension value and introduces the number of business pointers as the association expression dimension. This ensures that each association path not only reflects the business evolution status but also embodies the association carrying capacity after the continuous expansion of business pointers. This guarantees that each business pointer is mapped to its corresponding independent association entry point to form a continuous, consistent, and independent business association expression.

[0068] Based on the completed mapping of all associated paths, the association content between each unique semantic identifier is continuously organized to construct association path information. Association path information represents a complete description of the associations formed between each unique semantic identifier and its corresponding independent associated entry point, associated path, and business direction. It not only preserves the fixed correspondence between unique semantic identifiers and business directions but also records the continuous correspondence between the number of business scope changes, the number of referenced objects, the activity coefficient of the associated entry point, the extension value of the associated path, and the path association coverage coefficient. Therefore, it can fully express the independent association process formed during the continuous evolution of the field's business semantics. Each associated path in the association path information has a unique business source and a unique business destination. Different associated paths remain independent of each other, while multiple associated paths originating from the same field's evolutionary trajectory still retain the continuous connection of business development. When new business directions are added subsequently, only a new independent associated entry point needs to be established, and a new associated path needs to be generated according to the aforementioned relationship to supplement the association path information. The original associated paths remain continuously expressed, thus forming continuously expanding, uniformly organized association path information that accurately reflects the evolution of business semantics. This provides a complete data foundation for subsequently selecting corresponding associated paths based on the consistency of business directions.

[0069] To fully understand the specific implementation of this process, the following example further illustrates the details: Taking the power data field "customer_id" as an example, this field forms three business pointers in the field evolution trajectory, namely "marketing customer", "integrated energy customer" and "charging customer", with corresponding unique semantic identifiers of S001, S002 and S003 respectively. The number of times the business scope changes is 2, 4 and 6 times respectively, and the number of referenced objects changes is 3, 5 and 7 respectively.

[0070] First, the semantic identifier correspondence is read, and independent associated entry points E001, E002, and E003 are established for S001, S002, and S003 respectively. Each associated entry point corresponds to only one unique semantic identifier. Then, the activity coefficients of the associated entry points are calculated based on the number of business scope changes and the number of referenced objects, yielding R=2.40, R=4.17, and R=6.22 respectively. Next, the corresponding associated paths are generated by combining the number of business scope changes and the number of referenced objects, resulting in associated path extension values ​​L=14.40, L=41.70, and L=87.08 respectively, and associated paths P001, P002, and P003 are established accordingly. Then, the business targeting marketing customers is mapped to associated entry point E001 and associated path P001; the business targeting "integrated energy customers" is mapped to associated entry point E002 and associated path P002; and the business targeting charging customers is mapped to associated entry point E003 and associated path P002. The associated path P003 is defined as follows: P001 is associated with 2 business nodes, P002 with 4 business nodes, and P003 with 6 business nodes. The corresponding path association coverage coefficients are Q=2.88, Q=16.68, and Q=40.19, respectively. The final associated path information is constructed, including three complete association relationships: unique semantic identifier S001 - associated entry point E001 - associated path P001 - business pointing to marketing customers; unique semantic identifier S002 - associated entry point E002 - associated path P002 - business pointing to integrated energy customers; and unique semantic identifier S003 - associated entry point E003 - associated path P003 - business pointing to charging customers. This ensures that the same field corresponds to independent associated entry points and independent associated paths in different business scenarios, allowing direct location of the corresponding business path during subsequent semantic association retrieval, avoiding semantic confusion caused by multiple business points sharing the same associated path.

[0071] S004, Receive power data semantic association retrieval request, use association path information to match the fields in the query conditions to the corresponding unique identification mark, select the corresponding association path according to the degree of consistency of business direction, and determine the business association path;

[0072] In the semantic association retrieval process of power data, although different query conditions may correspond to the same field name, their actual business meanings are not consistent due to different business development directions. Therefore, it is necessary to identify the true business reference corresponding to the query conditions based on the association path information and select the association path consistent with the business semantics. Based on the already constructed association path information, the business semantics of the power data semantic association retrieval request is located to ultimately determine the business association path. The specific steps are as follows:

[0073] The system receives semantic association retrieval requests for power data and extracts field pointers from the query conditions. These field pointers are used as the core content for business semantic positioning and are correlated with the field sources in the associated path information. The field pointers in the query conditions include not only field names but also descriptive content that reflects business semantics, such as business descriptions, query objects, call sources, business scenarios, and associated objects. All of these elements together constitute the set of field pointer expressions.

[0074] When a query condition enters the retrieval process, the field names remain unchanged, while the field pointers present different business meanings depending on the different business scenarios. Therefore, the field pointer content corresponding to the query condition is recorded separately, and organized item by item according to the unique semantic identifier, unique identification mark, and associated path stored in the associated path information. This ensures that the field pointers in the query condition can enter the unified business semantic expression range, establish a complete business semantic foundation for subsequent matching of unique identification marks, and maintain the continuous correspondence between the query condition, field pointers, and associated path information.

[0075] Based on the fields pointed to in the query conditions, all unique identifiers stored in the associated path information are read. Then, according to the business description, call source, associated object, and business scope expression status corresponding to the field pointer, each field pointer is associated with the business semantics corresponding to its unique identifier, thus progressively mapping the field pointer to the unique identifier. Since each unique identifier in the associated path information corresponds to a unique semantic identifier and a unique business pointer, the same field may have multiple unique identifiers in multiple business scenarios. The field pointer in the query conditions is only associated with one of these unique identifiers; the remaining unique identifiers remain independent and do not participate in the current business semantic expression.

[0076] When a field pointer corresponds to a unique identifier, the associated path corresponding to that unique identifier simultaneously enters the candidate business association scope. This allows the business semantics in the query conditions to be gradually transformed from field names into unique business semantic expressions, while maintaining a one-to-one correspondence between field pointers and unique identifiers. This provides a unified business expression for subsequent business pointer consistency analysis.

[0077] Based on the unique identifiers that have established corresponding relationships, the business pointers in the corresponding associated paths are read, and the business semantic relationships between the field pointers in the query conditions and the business pointers in the associated paths are organized. A business pointer consistency description is established for each associated path. The business pointer consistency degree is used to indicate the business semantic compliance status between the field pointers in the query conditions and the corresponding business pointers in the associated paths. Its expression includes the business description correspondence status, the call source correspondence status, the business scope correspondence status, the associated object correspondence status, and the business development status. All business content together describe the business relationship between the field pointers and the business pointers.

[0078] When multiple related paths correspond to the same field, each related path forms an independent degree of consistency in business direction. Different business directions remain independent of each other, while the development relationship originating from the same field's evolutionary trajectory is retained. This allows each related path to have independent business semantics while maintaining the developmental connection formed by the continuous evolution of the field's business, thereby forming multiple candidate business related paths and providing a continuous and complete basis for determining subsequent business related paths.

[0079] Based on the degree of consistency in business orientation, a business association path that is consistent with the business semantics corresponding to the query conditions is selected from multiple candidate association paths, and then determined. A business association path represents the complete business association process formed by establishing a continuous business connection between the field pointer in the query conditions after it has been located by a unique identifier and the corresponding business pointer. It not only preserves the correspondence between the query conditions, field pointers, unique identifiers, and association paths, but also records the development process of business semantics from the query request entering the association path until the business positioning is completed. Therefore, the business association path can fully express the direction of business semantic flow corresponding to the query conditions. For the same field, different business scenarios correspond to different business association paths. Each business association path remains independent, while the business development relationship originating from the evolution trajectory of the same field is retained. This allows subsequent semantic association retrieval to directly conduct business data association along the determined business association path, without relying on field names for unified association. This ensures that different business scenarios enter the development path of their corresponding business semantics, achieving a continuous, consistent, and independent association expression between field business semantics and business association paths.

[0080] S005, Reorganize the field association content based on the business association path, extract the corresponding data content according to the business direction, and output the semantic association retrieval results of power data corresponding to the business association path;

[0081] The final stage of semantic association retrieval for power data uses the established business association path as the basis for business semantic transmission. Field association content is reorganized to ensure that the retrieval results always maintain a semantic expression consistent with the business intent, avoiding the sharing of the same field association content across different business scenarios. Based on the business association path, field association content is reorganized, corresponding data content is extracted according to the business intent, and the power data semantic association retrieval results corresponding to the business association path are output. The specific steps are as follows:

[0082] The system reads the established business association paths and organizes the field association content according to the business semantic development order stored in the business association paths. Field association content includes historical field attributes, business descriptions, call locations, associated objects, reference relationships, and business connections between fields. All content is organized based on the business pointers in the business association paths, rather than by field names. When a business association path corresponds to a specific business pointer, only the field association content consistent with that business pointer is retrieved, maintaining a continuous correspondence between field association content. This ensures that the field association content is arranged sequentially along the business association path, forming a complete business association expression.

[0083] At the same time, the same field has independent field-related content under different business association paths. Each business association path remains independent of the others, and the developmental connection originating from the same field's evolutionary trajectory is preserved. This enables the reorganization of field-related content from the field dimension to the business semantic dimension, establishing a unified business expression foundation for subsequent data content extraction.

[0084] Based on the reorganized field associations, the corresponding data content is organized item by item according to the business pointers in the business association path, ensuring that each business pointer in the path corresponds to its respective data content. The data content includes the actual stored content of the field, the associated business information, the business object to which the field belongs, the corresponding business record, and the business relationships between the associated data. All items maintain a consistent correspondence with their respective business pointers. When a business association path changes, the corresponding data content is adjusted synchronously. Different business association paths organize different data content. Even if the name of the same field remains the same, its corresponding data content remains independently expressed due to different business pointers, thus ensuring that the data content is always continuously organized around the business association path.

[0085] At the same time, the evolutionary relationship between the various data contents continues to be preserved from the same field, so that the development process of business semantics can be fully mapped to the corresponding data content, providing continuous data expression for the subsequent formation of complete search results.

[0086] Based on the extracted data, business correspondences are established between field-related content and data content. This ensures that the business descriptions, call locations, associated objects, and business relationships in the field-related content are consistently expressed in accordance with the business records in the data content, and this process is continuously organized along the business association path. At this point, the business association path not only serves as a semantic location for business information but also as a basis for continuous connection between field-related content and data content, enabling accurate mapping of field-related content to target data content. Different business association paths result in different business data representations.

[0087] When the same field corresponds to multiple business association paths, each business association path stores its corresponding data content separately, maintaining independent business semantics and avoiding cross-organization. However, the business development connections originating from the evolution trajectory of the same field are retained, so that the entire data organization process reflects the continuity of business semantic development while maintaining independent expression between different business scenarios. This forms a continuous and consistent business association relationship between field-related content, business orientation, and data content.

[0088] Based on the organized data content, output the semantic association retrieval results of power data corresponding to the business association path. The power data semantic association retrieval results represent the final business data expression formed after the query conditions have completed business semantic positioning along the determined business association path. It includes the business association path, the corresponding business focus, the associated field content, and the complete association relationship between the corresponding data content, maintaining a continuous state of business semantic development. For different query conditions, even if the same field name is involved, different power data semantic association retrieval results will be output as long as the corresponding business association path is different; for the same business association path, data content consistent with that business focus will be continuously output, ensuring that the retrieval results always maintain a business semantic expression consistent with the business association path.

[0089] Ultimately, the semantic association retrieval results of each power data point correspond to different business association paths, forming a set of business semantic retrieval results that are independent yet continuously organized. This provides unified, continuous, and accurate semantic association data support for subsequent power data queries, business analysis, and business semantic applications.

[0090] This application uses fields as the basic object of semantic evolution, continuously collecting evolutionary information such as historical field attributes, business descriptions, call locations, reference frequencies, and associated objects. It constructs field evolution trajectories, further pinpoints changes in business focus, completes semantic decomposition and establishes unique semantic identifiers. Subsequently, it establishes independent association entry points and corresponding association paths for different business semantics, completing the reorganization of association relationships based on the continuous evolution of field business semantics. When the same field gradually evolves into multiple business meanings, each business meaning has an independent semantic expression and association path. The retrieval process can enter the corresponding business association path based on different business focuses, no longer relying on the initially established single semantic association relationship. Therefore, it can maintain the synchronous evolution of field business semantics and association relationships, improve the adaptability of power data semantic association retrieval to business changes, and enhance the accuracy of semantic association results under different business scenarios.

[0091] This application reorganizes field association content around the established business association paths, extracts corresponding data content according to business directions, and finally outputs power data semantic association retrieval results corresponding to the business association paths, maintaining a continuous correspondence between the field business semantic positioning process and the data association process. When different business scenarios input the same field query conditions, the retrieval request can select the corresponding business association path based on the business direction and output data content matching the business semantics. Different business needs correspond to different retrieval results, avoiding semantic confusion caused by different business scenarios sharing the same field association relationship for a long time. This provides a consistent data association foundation for intelligent question answering, knowledge reasoning, and natural language query conversion, while enhancing the business differentiation capability of power data semantic association retrieval results and the credibility of intelligent data application processes.

[0092] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A semantic association retrieval method for power data based on field-level relationship evolution, characterized in that, Includes the following steps: S001, collect historical attributes, business descriptions, call locations, reference frequencies, and associated objects of power data fields, count the number of business pointer changes and the number of referenced objects changes, construct field evolution trajectories, and extract the change characteristics of multiple business pointers corresponding to the same field to generate field evolution trajectories; S002, locate the change position of each business pointer in the evolution trajectory of the positioning field, perform semantic decomposition on adjacent business pointers, generate a unique semantic identifier according to the difference of business pointers, configure a unique identification mark for each business pointer, and establish a semantic identifier correspondence; S003, Read the semantic identifier correspondence, establish an independent association entry for each unique semantic identifier, generate corresponding association paths according to the number of business scope changes and the number of referenced objects changes, map each business pointer to the corresponding association entry, and construct association path information; S004, Receive power data semantic association retrieval request, use association path information to match the fields in the query conditions to the corresponding unique identification mark, select the corresponding association path according to the degree of consistency of business direction, and determine the business association path; S005: Based on the business association path, reorganize the field association content, extract the corresponding data content according to the business direction, and output the semantic association retrieval results of power data corresponding to the business association path.

2. The power data semantic association retrieval method based on field-level relation evolution according to claim 1, characterized in that, In step S001, when collecting historical attributes, business descriptions, call locations, reference frequencies, and associated objects of power data fields, the historical records of the same field formed in different business stages are continuously organized according to the order of business occurrence. The historical attributes include field name, field type, data object to which the field belongs, field business description, and maintenance time. The number of business pointer changes is counted based on changes in business descriptions, and the number of referenced objects is counted based on changes in associated objects. The field evolution trajectory is constructed according to the continuous correspondence of the historical records. The field evolution trajectory is used to record the development process of the corresponding business pointer of the same field in multiple business stages and the continuous association between each business stage.

3. The power data semantic association retrieval method based on field-level relationship evolution according to claim 1, characterized in that, In step S002, when locating the change positions of each business direction in the field evolution trajectory, adjacent business stages in the field evolution trajectory are grouped into business direction change units, and the change positions of business directions are determined based on business descriptions, call locations, reference frequencies, and the corresponding relationships of associated objects. Semantic decomposition is performed for each change position of business direction, and the business semantics corresponding to the same field in different business stages are expressed independently. Unique semantic identifiers are generated according to the differences in business directions, so that different business directions correspond to different unique semantic identifiers.

4. The power data semantic association retrieval method based on field-level relationship evolution according to claim 3, characterized in that, In step S002, when configuring a unique identification mark for each business pointer, the unique identification mark is respectively mapped to a unique semantic identifier and a business pointer, and the fixed correspondence between the unique identification mark and the business description, call location, reference frequency and associated object is maintained; the semantic identifier correspondence is established based on all unique identification marks, so that when the same field corresponds to multiple business pointers, each business pointer corresponds to an independent unique semantic identifier and a unique identification mark, and the development relationship of each business pointer originating from the same field evolution trajectory is preserved.

5. The power data semantic association retrieval method based on field-level relationship evolution according to claim 4, characterized in that, In step S003, when establishing an independent association entry for each unique semantic identifier, each independent association entry corresponds to a unique semantic identifier and saves the corresponding business pointer, unique identification mark, number of business scope changes, and number of referenced object changes; corresponding association paths are generated based on the number of business scope changes and the number of referenced object changes, so that different unique semantic identifiers have independent association entry and independent association paths, and maintain the business development relationship originating from the same field evolution trajectory.

6. The power data semantic association retrieval method based on field-level relation evolution according to claim 5, characterized in that, In step S003, when mapping each business pointer to its corresponding associated entry point, a unique semantic identifier is used as the mapping basis, so that each business pointer corresponds to only one independent associated entry point. The business description, call location, associated object and business scope change information are organized along the corresponding associated path to construct the associated path information. The associated path information includes a unique semantic identifier, a unique identification mark, an independent associated entry point, an associated path and the correspondence between business pointers.

7. The power data semantic association retrieval method based on field-level relation evolution according to claim 6, characterized in that, In step S004, after receiving the power data semantic association retrieval request, the fields pointed to in the query conditions are extracted, and a correspondence is established with the unique identification mark in the association path information based on the business description, call source, business scenario and associated object corresponding to the field pointed to, so that the fields in the query conditions point to the corresponding unique business semantics and form the corresponding candidate association path set.

8. The power data semantic association retrieval method based on field-level relation evolution according to claim 7, characterized in that, In step S004, when selecting the corresponding associated path according to the degree of consistency of business direction, the business correspondence between the field direction in the query conditions and the corresponding business direction of each candidate associated path is established respectively. The degree of consistency of business direction is determined according to the corresponding status of business description, business scope, call source and associated object. The corresponding business associated path is then selected from the candidate associated paths.

9. The power data semantic association retrieval method based on field-level relation evolution according to claim 8, characterized in that, In step S005, when reorganizing the field association content based on the business association path, the historical attributes, business descriptions, call locations, associated objects, and business relationships between fields are reorganized according to the business pointers saved in the business association path. This ensures that the field association content is continuously organized along the business association path, with different business association paths corresponding to different field association content, while maintaining the business development relationship originating from the same field evolution trajectory.

10. The power data semantic association retrieval method based on field-level relation evolution according to claim 9, characterized in that, In step S005, when extracting the corresponding data content according to the business direction, the business records, business objects and related data content corresponding to the fields are extracted according to the business direction corresponding to the business association path, and the correspondence between the data content and the field association content is maintained; the semantic association retrieval results of power data corresponding to the business association path are output according to the extracted data content, so that different business association paths correspond to different semantic association retrieval results of power data.

Images