Multi-source heterogeneous ontology fusion method and device, and readable storage medium

Abstract

This application proposes a method, apparatus, and readable storage medium for multi-source heterogeneous ontology fusion. The method includes: acquiring structured metadata from multiple heterogeneous physical data tables, and ontology structured data from at least two sets of heterogeneous ontology standards; constructing an intermediate semantic layer based on the structured metadata and ontology structured data; mapping each heterogeneous physical data table to the intermediate semantic layer to generate a first mapping relationship dataset; mapping the intermediate semantic layer to each heterogeneous ontology standard to generate a second mapping relationship dataset; constructing a unified knowledge graph ontology library based on the first and second mapping relationship datasets; determining the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library; routing to the corresponding heterogeneous physical data table according to the mapping path; performing field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard, and outputting structured retrieval result data. This application reduces mapping complexity and achieves ontology fusion.

Description

Technical Field

[0001] This invention relates to the field of knowledge graph and data processing technology, and more specifically, to a method, apparatus and readable storage medium for multi-source heterogeneous ontology fusion. Background Technology

[0002] In related technologies, the common multi-source heterogeneous ontology fusion often adopts a point-to-point mapping method, which results in an exponential increase in mapping complexity. The physical data table is strongly bound to the ontology standard, making it impossible to be compatible with multiple heterogeneous ontology standards in a unified system. The maintenance cost of mapping rules is high, and cross-standard data interoperability is difficult. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art or related art.

[0004] Therefore, the first aspect of the present invention proposes a multi-source heterogeneous ontology fusion method.

[0005] A second aspect of the present invention provides a multi-source heterogeneous body fusion device.

[0006] A third aspect of the present invention proposes another multi-source heterogeneous body fusion device.

[0007] The fourth aspect of this application proposes a readable storage medium.

[0008] In view of this, a first aspect of the present invention provides a multi-source heterogeneous ontology fusion method, comprising: acquiring structured metadata of multiple heterogeneous physical data tables and ontology structured data of at least two sets of heterogeneous ontology standards; constructing an intermediate semantic layer independent of the heterogeneous physical data tables and heterogeneous ontology standards based on the structured metadata and ontology structured data; mapping each heterogeneous physical data table to the intermediate semantic layer to generate a first mapping relationship dataset; mapping the intermediate semantic layer to each heterogeneous ontology standard to generate a second mapping relationship dataset; constructing a unified knowledge graph ontology library based on the first mapping relationship dataset and the second mapping relationship dataset; responding to a cross-ontology standard data retrieval instruction, determining the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library; routing to the corresponding heterogeneous physical data table according to the mapping path and extracting source data from the heterogeneous physical data table; performing field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard, and outputting structured retrieval result data.

[0009] This application provides a multi-source heterogeneous ontology fusion method, which first involves acquiring structured metadata from multiple heterogeneous physical data tables and ontology structured data from at least two heterogeneous ontology standards. Heterogeneous physical data tables refer to database tables originating from different business systems, such as Enterprise Resource Planning (ERP), Customer Relationship Management (CRM), and Human Resource Management (HR). The table names, field names, data types, primary and foreign key relationships, and inter-table relationships constitute the structured metadata. Heterogeneous ontology standards refer to ontology libraries from different systems, such as enterprise-internal proprietary standards (e.g., A semantic model) and international industry standards (e.g., Financial Industry Business Ontology, FIBO). The ontology structured data includes ontology classes, attributes, inter-class relationships, and identification information such as Uniform Resource Identifiers (URIs). Acquiring these two types of data lays the data foundation for subsequent semantic fusion. Next, based on the structured metadata and ontology structured data, an intermediate semantic layer independent of the heterogeneous physical data tables and heterogeneous ontology standards is constructed. The intermediate semantic layer is an enterprise-level intermediate ontology independent of any specific business system and external standards. Using `ent` as its namespace (i.e., the enterprise-level ontology namespace), it contains only the core business concepts of the enterprise, such as `ent colon Person` for a natural person and `ent colon Order` for an order, serving as the hub for semantic transformation. Then, each heterogeneous physical data table is mapped to the intermediate semantic layer to generate the first mapping relationship dataset. This step uses semantic matching rules to map each heterogeneous physical data table, such as the order table named `tunderscoreorder`, to its corresponding semantic entity in the intermediate semantic layer, such as `ent colon Order`, thus generating a set of mapping relationships between the physical data layer and the intermediate semantic layer, denoted as the first mapping relationship dataset. Next, the intermediate semantic layer is mapped to various heterogeneous ontology standards to generate the second mapping relationship dataset. This involves utilizing subclass relations (RDF) or equivalent class relations (OWL) in Web ontology languages, such as the subclass of an entity (rdfs) or equivalent class of an entity (owl), to connect semantic entities in the intermediate semantic layer (e.g., ent) to enterprise-internal proprietary standards (e.g., yonyou) and international industry standards (e.g., fibo). This generates a set of mapping relationships between the intermediate semantic layer and various heterogeneous ontology standards, denoted as the second mapping relationship dataset. Subsequently, a unified knowledge graph ontology is constructed based on the first and second mapping relationship datasets. The unified knowledge graph ontology is a complete semantic network containing physical data tables, intermediate semantic layer entities, and multiple sets of external standard ontology classes, with all nodes interconnected through mapping relationships.When the system receives a cross-ontology standard data retrieval command, such as a user requesting to query loan contract information according to the FIBO standard, it determines the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library. The mapping path refers to the complete link starting from the target heterogeneous ontology standard, such as FIBO, finding the intermediate semantic layer entity along the second mapping relationship dataset, and then finding the corresponding heterogeneous physical data table along the first mapping relationship dataset. Subsequently, the system routes to the corresponding heterogeneous physical data table according to the mapping path and extracts source data from that table. The system directly locates the specific physical table, such as the `t underscore loan underscore agmt` table, based on the mapping path and extracts the original records. Finally, the system performs field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard, outputting structured retrieval results. That is, based on the field format, attribute naming, and relation structure of the target ontology standard, such as FIBO, the extracted source data undergoes field mapping, data type conversion, and relation reconstruction to generate structured data such as Resource Description Framework triples (RDF triples) or JSON linked data (JSON-LD) conforming to the standard format, and returns it to the caller.

[0010] The above steps introduce an intermediate semantic layer as a semantic transformation hub, splitting the original direct mapping from the physical data layer to multiple heterogeneous ontology standards into two stages: from the physical layer to the intermediate layer and from the intermediate layer to multiple standards. Since the first mapping dataset and the second mapping dataset are independent, adding a new external standard only requires supplementing the mapping from the intermediate layer to the new standard; there is no need to modify the existing mapping rules from the physical layer to the intermediate layer. Therefore, the total maintenance quantity of mapping rules is reduced from the multiplicative level (N x M) of the traditional point-to-point approach to the additive level (N + M), where N is the number of physical tables and M is the number of standard sets. This fundamental architectural change brings about low coupling and high scalability, enabling the system to support any number of heterogeneous ontology standards with linearly increasing maintenance costs, while simultaneously achieving semantic interoperability of the same physical data in different standard contexts.

[0011] In some technical solutions of this application, the structured metadata of heterogeneous physical data tables includes at least one of table name, field name, field type, primary key relationship, foreign key relationship and inter-table association relationship; the ontology structured data of heterogeneous ontology standards includes at least one of ontology class, attribute, inter-class relationship and identification information.

[0012] In some technical solutions of this application, constructing an intermediate semantic layer independent of heterogeneous physical data tables and heterogeneous ontology standards includes: configuring a unified semantic root node and determining that the semantic classes mapped from each heterogeneous physical data table are all subclasses of the unified semantic root node; setting at least one basic semantic axis under the unified semantic root node, and generating structured definition data of semantic entities and semantic relationships based on the basic semantic axis.

[0013] In some technical solutions of this application, the basic semantic axis includes at least one of the following: document semantic axis, economic semantic axis, process semantic axis, subject semantic axis, rule semantic axis, master data semantic axis, and detail semantic axis.

[0014] In some technical solutions of this application, a composite semantic axis is generated based on at least two basic semantic axes, and the corresponding heterogeneous physical data table is mapped to the basic semantic axis or the composite semantic axis according to the table characteristics and field characteristics of the heterogeneous physical data table.

[0015] In some technical solutions of this application, the intermediate semantic layer is mapped to various heterogeneous ontology standards to generate a second mapping relationship dataset, including: based on the matching results of semantic entities and ontology classes, attributes and inter-class relationships in various heterogeneous ontology standards, a cross-standard mapping relationship is established according to ontology subclass relationships and / or ontology equivalence class relationships.

[0016] In some technical solutions of this application, the multi-source heterogeneous ontology fusion method also includes a processing step for adding a new heterogeneous ontology standard. The processing steps include: obtaining the new heterogeneous ontology standard file and parsing it to obtain the new ontology structured data; configuring a global priority score for each heterogeneous ontology standard; when the identification information and name of the new ontology class conflict with the semantic entity in the intermediate semantic layer, determining the conflict handling result according to the corresponding global priority score; wherein, when adding a new heterogeneous ontology standard, the first mapping relationship dataset remains unchanged; and completing the fusion deduplication and updating the unified knowledge graph ontology library according to the conflict handling result.

[0017] A second aspect of the present invention provides a multi-source heterogeneous ontology fusion device, comprising: a first acquisition module, a first determination module, a second determination module, a third determination module, a fourth determination module, a fifth determination module, a sixth determination module, and a seventh determination module; the first acquisition module is used to acquire structured metadata of multiple heterogeneous physical data tables, and ontology structured data of at least two sets of heterogeneous ontology standards; the first determination module is used to construct an intermediate semantic layer independent of the heterogeneous physical data tables and heterogeneous ontology standards based on the structured metadata and ontology structured data; the second determination module is used to map each heterogeneous physical data table to the intermediate semantic layer to generate a first mapping relationship dataset; The third determining module is used to map the intermediate semantic layer to various heterogeneous ontology standards to generate a second mapping relation dataset; the fourth determining module is used to construct a unified knowledge graph ontology library based on the first and second mapping relation datasets; the fifth determining module is used to determine the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library in response to cross-ontology standard data retrieval instructions; the sixth determining module is used to route to the corresponding heterogeneous physical data table according to the mapping path and extract source data from the heterogeneous physical data table; the seventh determining module is used to perform field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard and output structured retrieval result data.

[0018] A third aspect of the present invention provides a multi-source heterogeneous ontology fusion apparatus, comprising: a processor and a memory, wherein the memory stores a program or instructions, and the processor, when executing the program or instructions in the memory, implements the steps of the multi-source heterogeneous ontology fusion method as described in any of the above-described technical solutions. Therefore, the multi-source heterogeneous ontology fusion apparatus possesses all the beneficial effects of the multi-source heterogeneous ontology fusion method as described in any of the above-described technical solutions.

[0019] A fourth aspect of the present invention provides a readable storage medium storing a program or instructions, which, when executed by a processor, implement the steps of the multi-source heterogeneous ontology fusion method as described in any of the above-described technical solutions. Therefore, the readable storage medium possesses all the beneficial effects of the multi-source heterogeneous ontology fusion method as described in any of the above-described technical solutions.

[0020] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0022] Figure 1 This is one of the flowcharts illustrating a multi-source heterogeneous ontology fusion method according to an embodiment of the present invention;

[0023] Figure 2 This is a logical diagram of the three-layer bridging architecture topology of a multi-source heterogeneous ontology fusion method according to an embodiment of the present invention;

[0024] Figure 3 A simulation diagram for the semantic interoperability and conflict resolution experiment of a multi-source heterogeneous ontology fusion method according to an embodiment of the present invention;

[0025] Figure 4 This is one of the schematic block diagrams of a multi-source heterogeneous body fusion device according to an embodiment of the present invention;

[0026] Figure 5 This is a second schematic block diagram of a multi-source heterogeneous body fusion device according to an embodiment of the present invention. Detailed Implementation

[0027] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0028] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0029] The following reference Figures 1 to 5 This invention describes a multi-source heterogeneous ontology fusion method, apparatus, and readable storage medium according to some embodiments of the present invention.

[0030] like Figure 1 As shown, embodiments of this application provide a multi-source heterogeneous ontology fusion method, the steps of which include:

[0031] Step 102: Obtain structured metadata from multiple heterogeneous physical data tables, as well as ontology structured data from at least two heterogeneous ontology standards;

[0032] Step 104: Based on structured metadata and ontology structured data, construct an intermediate semantic layer that is independent of heterogeneous physical data tables and heterogeneous ontology standards.

[0033] Step 106: Map each heterogeneous physical data table to the intermediate semantic layer to generate the first mapping relationship dataset;

[0034] Step 108: Map the intermediate semantic layer to each heterogeneous ontology standard to generate a second mapping relation dataset;

[0035] Step 110: Construct a unified knowledge graph ontology based on the first mapping relationship dataset and the second mapping relationship dataset;

[0036] Step 112: In response to the cross-ontology standard data retrieval command, determine the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library;

[0037] Step 114: Route to the corresponding heterogeneous physical data table according to the mapping path, and extract source data from the heterogeneous physical data table;

[0038] Step 116: Perform field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard, and output structured search result data.

[0039] This application provides a multi-source heterogeneous ontology fusion method, which first involves acquiring structured metadata from multiple heterogeneous physical data tables and ontology structured data from at least two heterogeneous ontology standards. Heterogeneous physical data tables refer to database tables originating from different business systems, such as Enterprise Resource Planning (ERP), Customer Relationship Management (CRM), and Human Resource Management (HR). The table names, field names, data types, primary and foreign key relationships, and inter-table relationships constitute the structured metadata. Heterogeneous ontology standards refer to ontology libraries from different systems, such as enterprise-internal proprietary standards (e.g., A semantic model) and international industry standards (e.g., Financial Industry Business Ontology, FIBO). The ontology structured data includes ontology classes, attributes, inter-class relationships, and identification information such as Uniform Resource Identifiers (URIs). Acquiring these two types of data lays the data foundation for subsequent semantic fusion. Next, based on the structured metadata and ontology structured data, an intermediate semantic layer independent of the heterogeneous physical data tables and heterogeneous ontology standards is constructed. The intermediate semantic layer is an enterprise-level intermediate ontology independent of any specific business system and external standards. Using `ent` as its namespace (i.e., the enterprise-level ontology namespace), it contains only the core business concepts of the enterprise, such as `ent colon Person` for a natural person and `ent colon Order` for an order, serving as the hub for semantic transformation. Then, each heterogeneous physical data table is mapped to the intermediate semantic layer to generate the first mapping relationship dataset. This step uses semantic matching rules to map each heterogeneous physical data table, such as the order table named `tunderscoreorder`, to its corresponding semantic entity in the intermediate semantic layer, such as `ent colon Order`, thus generating a set of mapping relationships between the physical data layer and the intermediate semantic layer, denoted as the first mapping relationship dataset. Next, the intermediate semantic layer is mapped to various heterogeneous ontology standards to generate the second mapping relationship dataset. This involves utilizing subclass relations (RDF) or equivalent class relations (OWL) in Web ontology languages, such as the subclass of an entity (rdfs) or equivalent class of an entity (owl), to connect semantic entities in the intermediate semantic layer (e.g., ent) to enterprise-internal proprietary standards (e.g., yonyou) and international industry standards (e.g., fibo). This generates a set of mapping relationships between the intermediate semantic layer and various heterogeneous ontology standards, denoted as the second mapping relationship dataset. Subsequently, a unified knowledge graph ontology is constructed based on the first and second mapping relationship datasets. The unified knowledge graph ontology is a complete semantic network containing physical data tables, intermediate semantic layer entities, and multiple sets of external standard ontology classes, with all nodes interconnected through mapping relationships.When the system receives a cross-ontology standard data retrieval command, such as a user requesting to query loan contract information according to the FIBO standard, it determines the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library. The mapping path refers to the complete link starting from the target heterogeneous ontology standard, such as FIBO, finding the intermediate semantic layer entity along the second mapping relationship dataset, and then finding the corresponding heterogeneous physical data table along the first mapping relationship dataset. Subsequently, the system routes to the corresponding heterogeneous physical data table according to the mapping path and extracts source data from that table. The system directly locates the specific physical table, such as the `t underscore loan underscore agmt` table, based on the mapping path and extracts the original records. Finally, the system performs field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard, outputting structured retrieval results. That is, based on the field format, attribute naming, and relation structure of the target ontology standard, such as FIBO, the extracted source data undergoes field mapping, data type conversion, and relation reconstruction to generate structured data such as Resource Description Framework triples (RDF triples) or JSON linked data (JSON-LD) conforming to the standard format, and returns it to the caller.

[0040] The above steps introduce an intermediate semantic layer as a semantic transformation hub, splitting the original direct mapping from the physical data layer to multiple heterogeneous ontology standards into two stages: from the physical layer to the intermediate layer and from the intermediate layer to multiple standards. Since the first mapping dataset and the second mapping dataset are independent, adding a new external standard only requires supplementing the mapping from the intermediate layer to the new standard; there is no need to modify the existing mapping rules from the physical layer to the intermediate layer. Therefore, the total maintenance quantity of mapping rules is reduced from the multiplicative level (N x M) of the traditional point-to-point approach to the additive level (N + M), where N is the number of physical tables and M is the number of standard sets. This fundamental architectural change brings about low coupling and high scalability, enabling the system to support any number of heterogeneous ontology standards with linearly increasing maintenance costs, while simultaneously achieving semantic interoperability of the same physical data in different standard contexts.

[0041] In some embodiments of this application, the structured metadata of heterogeneous physical data tables includes at least one of table name, field name, field type, primary key relationship, foreign key relationship, and inter-table association relationship; the ontology structured data of heterogeneous ontology standards includes at least one of ontology class, attribute, inter-class relationship, and identification information.

[0042] In the above embodiments, the specific content of structured metadata and ontology structured data is further defined. Specifically, the structured metadata of heterogeneous physical data tables includes at least one of the following: table name, field name, field type, primary key relationship, foreign key relationship, and inter-table association. Specifically, a table name, such as `t underscore proj underscore mgrunderscore ext`, uniquely identifies the physical table; a field name, such as `createunderscoretime`, identifies a column in the table. Field types, such as `VARCHAR` (variable-length string) and `INT` (integer), constrain the format of the data stored in the field. A primary key relationship identifies one or more columns that uniquely identify each row in the table, such as the order number column in the order table. A foreign key relationship identifies one or more columns in the table that reference the primary key of another table, such as the customer number column in the order table referencing the primary key in the customer table. Inter-table associations include one-to-many relationships, many-to-many relationships, and other connection logic, such as a customer having multiple orders. This metadata is the basic input data for subsequent semantic matching algorithms; the system can reconstruct the physical structure of the database by parsing this metadata. The ontology structured data of the heterogeneous ontology standard includes at least one of the following: ontology class, attributes, inter-class relationships, and identification information. An ontology class, such as `fibo.LoanAgreement`, represents the concept of a loan agreement. Attributes, such as `fibo.hasBorrower`, represent the relationship between the lender and borrower. Class relationships, such as `rdfs.subClassOf`, represent the inheritance relationship between classes. Identification information, such as a Uniform Resource Identifier (URI), is used to uniquely identify an ontology class globally, ensuring that ontology classes from different sources are not confused. By clearly defining the specific components of the above data structure, this method can accurately parse heterogeneous physical tables from different data sources and heterogeneous ontology libraries from different standard organizations, thereby supporting subsequent automated mapping and fusion, improving the accuracy of semantic matching and the configurability of the system, and avoiding mapping errors caused by ambiguous data formats.

[0043] In some embodiments of this application, constructing an intermediate semantic layer independent of heterogeneous physical data tables and heterogeneous ontology standards includes: configuring a unified semantic root node and determining that the semantic classes mapped from each heterogeneous physical data table are all subclasses of the unified semantic root node; setting at least one basic semantic axis under the unified semantic root node, and generating structured definition data of semantic entities and semantic relationships based on the basic semantic axis.

[0044] The above embodiments detail the specific implementation of constructing the intermediate semantic layer. First, a unified semantic root node is configured, and it is determined that all semantic classes mapped from heterogeneous physical data tables are subclasses of the unified semantic root node. The unified semantic root node is the top-level abstraction of the intermediate semantic layer, for example, defined as `ent_colon_BusinessObject`, where `BusinessObject` represents an enterprise business object, serving as the unique root class for all semantic classes. Any semantic class generated from mapping from heterogeneous physical data tables, such as `ent_colon_Order` for orders and `ent_colon_Person` for natural persons, must be declared as a subclass of `ent_colon_BusinessObject`. This specification provides an absolutely unified semantic anchor point in a multi-source heterogeneous environment, ensuring that data nodes, regardless of the isolated system they access, can ultimately connect to the same graph network, thus avoiding semantic silos. Second, at least one basic semantic axis is set under the unified semantic root node, and structured definition data of semantic entities and semantic relationships are generated based on this basic semantic axis. The basic semantic axes are gene-based classifications of the essence of business operations. Examples include DocumentArchetype, EconomicArchetype, ProcessArchetype, AgentArchetype, RuleArchetype, MasterDataArchetype, and EntryArchetype. Each basic semantic axis represents a core business prototype. For instance, the DocumentArchetype represents business vouchers such as expense reports and loan applications, while the EconomicArchetype represents economic activities such as expenses and revenue. The system maps physical tables to the corresponding basic semantic axes based on their name and field characteristics. For example, tables with the word "expense" in their name tend to be categorized under the EconomicArchetype. It then generates specific semantic entities based on these axes, such as "ent" followed by a colon to indicate an expense report (e.g., "ExpenseBill" for expense reports) and semantic relationships between entities (e.g., "ent" followed by a colon to indicate a submission relationship). With the above configuration, the intermediate semantic layer covers tens of thousands of complex physical tables of an enterprise with a minimal semantic skeleton, achieving efficient convergence from physical tables to abstract semantics, thus providing a stable and easily extensible intermediate representation for subsequent cross-standard mapping.

[0045] In some embodiments of this application, the basic semantic axis includes at least one of the following: document semantic axis, economic semantic axis, process semantic axis, subject semantic axis, rule semantic axis, master data semantic axis, and detail semantic axis.

[0046] The above embodiments specifically list the optional types of basic semantic axes. Basic semantic axes include at least one of the following: document semantic axis, economic semantic axis, process semantic axis, subject semantic axis, rule semantic axis, master data semantic axis, and detail semantic axis. Specifically, the document semantic axis (DocumentArchetype) is used to represent business voucher entities such as expense reports, loan applications, and orders. These entities typically include fields such as document number, document creation date, and amount. The economic semantic axis (EconomicArchetype) is used to represent entities related to economic activities such as expenses, revenue, and costs. These entities often involve fund flows and accounting. The process semantic axis (ProcessArchetype) is used to represent process control entities such as approval flows, workflows, and change processes. These entities typically include fields such as approval nodes, approvers, and approval status. The subject semantic axis (AgentArchetype) is used to represent behavioral subject entities such as natural persons, organizations, and departments. These entities are the initiators or participants in business activities. The rule semantic axis (RuleArchetype) is used to represent constraint rule entities such as reimbursement rules, credit rules, and approval rules. These entities define the judgment conditions of business logic. The MasterDataArchetype represents basic master data entities such as materials, products, customers, and suppliers. These entities are relatively stable and shared across business operations. The EntryArchetype represents detailed entities such as entry rows, detail rows, and sub-table rows. These entities typically serve as subordinate records under the master table entity. By pre-defining these semantic axes, the system can provide standardized classification metrics for heterogeneous physical tables. During the mapping process, the system automatically identifies the table name keywords and field characteristics of the physical table and assigns the physical table the corresponding prototype gene. For example, a table containing the keyword "expense" will be automatically marked as an economic axis entity. This semantic axis-based classification method explicitly expresses the semantic features of the underlying physical tables. Therefore, when mapping upwards to external international standards such as FIBO, the system can quickly locate the corresponding conceptual category in the standard based on the type of semantic axis, significantly reducing semantic ambiguity and improving the accuracy and consistency of the mapping.

[0047] In some embodiments of this application, a composite semantic axis is generated based on at least two basic semantic axes, and the corresponding heterogeneous physical data table is mapped to the basic semantic axis or the composite semantic axis according to the table characteristics and field characteristics of the heterogeneous physical data table.

[0048] In the above embodiments, a composite semantic axis and a mapping mechanism based on table features and field features are further introduced. Specifically, a composite semantic axis is generated based on at least two basic semantic axes. For example, the process semantic axis and the document semantic axis are combined through a multiple inheritance mechanism to generate a composite semantic axis, namely the Application axis (ApplicationArchetype). This axis is used to represent application-type businesses that have both process attributes and document attributes. For example, an expense application form requires both an approval process and a document structure. Similarly, the economic semantic axis is combined with the application axis again to generate the Expense axis (ExpenseArchetype), which is used to represent composite semantic scenarios such as expense reimbursement that have economic meaning and include both process and document attributes. Through this cross-axis multiple inheritance mechanism, a single physical table can simultaneously obtain semantic tags from multiple dimensions, thus possessing rich composite semantics. For example, an expense reimbursement form can be regarded as an economic axis entity for financial accounting, a process axis entity for approval tracking, and a document axis entity for voucher printing. Secondly, according to the table features and field features of the heterogeneous physical data tables, the corresponding heterogeneous physical data tables are mapped to basic semantic axes or composite semantic axes. Table features include keywords in the table name, such as "expense," and business descriptions in the table comments; field features include field names, such as "approverunderscoreid," which indicates the approver's number and implies a process axis, and field types, such as date fields, which imply time-related semantics. The system automatically identifies these features and matches the most suitable basic semantic axis or composite semantic axis for each physical table without manual intervention. For example, a table named "texpensereport" containing the fields "approverid" and "amount" will be automatically matched to the composite semantic axis of "expense." The underlying physical tables in this application acquire composite semantics in the intermediate semantic layer through a multi-axis inheritance mechanism, enabling a single physical table to exhibit different semantic aspects in different business contexts. This allows for precise matching of the corresponding standard concepts based on the characteristics of each standard when mapping upwards to multiple sets of external standards. For example, when mapping to financial standards, the economic axis semantics are emphasized, and when mapping to process standards, the process axis semantics are emphasized. This mechanism further reduces the complexity and ambiguity of mapping, significantly improves the accuracy of semantic interoperability, and gives the intermediate semantic layer the polymorphic characteristics of object-oriented languages.

[0049] In some embodiments of this application, the intermediate semantic layer is mapped to various heterogeneous ontology standards to generate a second mapping relationship dataset, including: establishing cross-standard mapping relationships based on the matching results of semantic entities and ontology classes, attributes and inter-class relationships in various heterogeneous ontology standards, according to ontology subclass relationships and / or ontology equivalence class relationships.

[0050] In the above embodiments, this application defines the specific method for mapping the intermediate semantic layer to various heterogeneous ontology standards. Based on the matching results between semantic entities and ontology classes, attributes, and inter-class relationships in each heterogeneous ontology standard, the system calculates semantic similarity and performs structural matching between the semantic entities generated in the intermediate semantic layer (e.g., ent_colon_Party representing a participant) and the ontology classes in each heterogeneous ontology standard. These heterogeneous ontology standards include internal enterprise proprietary standards such as yonyou_colon_Party, and international industry standards such as fibo_colon_LegalEntity representing a legal entity. The matching criteria include the string similarity of entity names, the overlap of attribute sets possessed by the entity, and the degree of matching of the semantic axis type of the entity. Then, cross-standard mapping relationships are established according to ontology subclass relationships and / or ontology equivalence class relationships. Specifically, if the semantic entities in the intermediate semantic layer are semantically completely equivalent to the ontology classes in the external standard, that is, they represent the same business concept and have the same attributes and behaviors, then an equivalence class relationship (owl_colon_equivalentClass) is established. If the semantic entities in the intermediate semantic layer are more concrete or abstract concepts of the external standard ontology class, then a subclass relationship (RDFS colon subClassOf) or a superclass relationship (RDFS colon superClassOf) is established. For example, `ent colon LoanContract` means that a loan contract can be declared as a subclass of `fibo colon LoanAgreement`, and also as an equivalent class of `yonyou colon Contract`. Through the above mapping method, the intermediate semantic layer becomes a bridge connecting multiple heterogeneous ontology standards. Since all physical data tables have been uniformly mapped to the intermediate semantic layer through the first mapping relationship dataset, only one mapping needs to be established between the intermediate semantic layer and each standard to enable all physical data tables to simultaneously support queries and inferences from multiple standards. This radial topology avoids establishing separate point-to-point mappings for each pair of physical tables and each standard, thereby reducing the number of mapping rules from multiplicative to additive, greatly reducing manual maintenance costs, and ensuring semantic consistency.

[0051] In some embodiments of this application, the multi-source heterogeneous ontology fusion method further includes a processing step for adding a new heterogeneous ontology standard. The processing step includes: obtaining the new heterogeneous ontology standard file and parsing it to obtain the new ontology structured data; configuring a global priority score for each heterogeneous ontology standard; when the identification information and name of the new ontology class conflict with the semantic entity in the intermediate semantic layer, determining the conflict handling result according to the corresponding global priority score; wherein, when adding a new heterogeneous ontology standard, the first mapping relationship dataset remains unchanged; and completing the fusion deduplication and updating the unified knowledge graph ontology library according to the conflict handling result.

[0052] In the above embodiments, this application provides dynamic processing steps for adding heterogeneous ontology standards, ensuring that the system has plug-and-play extensibility. First, the system acquires the new heterogeneous ontology standard file and parses it to obtain the new ontology's structured data. For example, when an enterprise needs to access the internationally recognized financial industry ontology FIBO, the system automatically reads FIBO's OWL (Web Ontology Language) file, parses out all ontology classes (e.g., FIBO colon LoanAgreement), attributes (e.g., FIBO colon hasPrincipal representing principal), and inter-class relationships (e.g., subclass inheritance). Second, a global priority score is configured for each heterogeneous ontology standard. The system allows administrators to set priority values ​​for different standards; for example, the priority of the enterprise's internal core standard is 100, the FIBO industry standard is 95, and the basic general standard BFO (Basic Form Ontology) is 90. The priority score can be manually adjusted according to the enterprise's actual business needs; for example, the score of the internal standard can be increased when the enterprise needs to prioritize internal standards. Then, when the identification information and name of the newly added ontology class conflict with those of semantic entities in the intermediate semantic layer, the conflict resolution result is determined based on the corresponding global priority score. The identification information refers to the Uniform Resource Identifier (URI) of the ontology, and the name refers to the local name of the ontology class. If the newly added FIBO colon Party has the exact same URI and name as the existing ent colon Party in the intermediate semantic layer, it is considered a conflict. In this case, the system compares the global priority scores of the two. If the score is higher, for example, 100 points for the internal enterprise standard is higher than 95 points for the FIBO standard, the higher-scoring standard automatically overrides the metadata definition of the lower-scoring standard. That is, the ent colon Party is retained as the main class, while the corresponding class of the lower-scoring standard, such as FIBO, is attached as a supplement to the main class of the higher-scoring standard. The key limitation is that when adding a heterogeneous ontology standard, the first mapping relationship dataset remains unchanged. This means that adding a new set of external standards does not require modifying the existing mapping rules from the physical data table to the intermediate semantic layer, because the first mapping relationship dataset only involves the mapping between the physical table and the ent layer, and is unrelated to the new standard. This fully reflects the low coupling characteristic of this invention, enabling the system to dynamically expand support for any number of new standards without interrupting existing services or reconstructing existing mappings. Finally, based on the conflict handling results, the system completes the fusion, deduplication, and updates the unified knowledge graph ontology library. Based on the priority decision, the system retains the main class of the high-scoring standard as an active node in the unified knowledge graph ontology, and mounts all attribute definitions of the low-scoring standard, such as attribute names, data types, and relationships, such as object attributes, onto the high-scoring main class to form supplementary metadata. Then, the graph library is reloaded, so that the new standard and its integration with existing nodes take effect immediately.Through the dynamic expansion steps of this application, the system supports plug-and-play functionality for any number of external standards without downtime, hard-coded mapping rules, or mapping table reconstruction. It automatically completes the dynamic reorganization and deduplication of the semantic web, thereby achieving extremely high system scalability and operational efficiency. This enables enterprises to flexibly access new industry standards or regulatory standards as their business develops.

[0053] In one embodiment, the system of this application also provides a pre-built and automatic mapping function for external standard nodes. Specifically, for ontology classes in external ontology standards that do not currently have corresponding physical data tables, the system does not directly ignore them, but pre-loads these external standard nodes into the unified knowledge graph ontology library. For example, when accessing the FIBO standard, even if no physical table in the system can be mapped to the FIBO "CreditRating" ontology class, the system still creates the node in the unified knowledge graph ontology library. At the same time, the system configures the initial state of the external standard node as a pending mapping state. When new heterogeneous physical data tables are subsequently accessed, the system extracts the table name of the physical table, performs normalization processing, including removing meaningless prefixes such as "tunderscore" and "sysunderscore", and performs word segmentation according to camelCase or underscore naming conventions. If the normalized table name segmentation result matches the name of a pre-loaded external standard node, for example, if the segmentation result has the same stem as the node name, the system automatically generates a corresponding semantic entity in the intermediate semantic layer. The namespace of this semantic entity is 'ent,' and its name maintains semantic consistency with the external standard node name. Then, the system automatically establishes an equivalence class or subclass relationship between this newly generated intermediate semantic entity and the pre-loaded external standard node, and simultaneously establishes the first mapping relationship between this intermediate semantic entity and the heterogeneous physical data table. Through this pre-defined and same-name matching mechanism, the system ensures that no external standard node is isolated from the semantic network, and that any subsequently connected physical data node can automatically find its corresponding semantic anchor point. This achieves progressive and fully automated knowledge graph construction, further improving the graph's coverage and construction efficiency. This embodiment is particularly suitable for scenarios where large enterprises frequently access new business systems and new industry standards, significantly reducing manual intervention costs and achieving true zero-configuration dynamic expansion.

[0054] like Figure 2 and Figure 3 As shown, Figure 2 The schematic diagram of the three-layer architecture for multi-source heterogeneous ontology fusion is an overall architectural diagram of the core technical solution of this invention. It fully presents the three-layer radial ontology fusion architecture with the intermediate semantic layer as the core hub, and intuitively reflects the core concept of this invention to solve the problems of high complexity of point-to-point mapping and strong binding between physical data and ontology standards in the prior art.

[0055] The bottom layer is the physical data source layer, which is the basic data source for semantic fusion in this invention. It contains multiple sets of heterogeneous physical data tables from different business systems with different structures and naming rules, covering core business tables such as user tables, personnel tables, sales order tables, and loan contract tables. All data tables are connected to the intermediate semantic layer through mapping relationships, which is the core processing object of this invention.

[0056] The middle layer is the enterprise core semantic layer (ent layer), which is the core innovative carrier that distinguishes this invention from existing technologies, and also the semantic conversion hub connecting the underlying data and the upper-layer standards. This layer is completely independent of the underlying physical data source and the upper-layer heterogeneous ontology standards, and constructs a unified standardized semantic entity, which serves as the "greatest common denominator of semantics" between the underlying heterogeneous data and the multiple sets of upper-layer ontology standards, realizing the unified semantic convergence of the underlying heterogeneous data.

[0057] The upper layer is a multi-standard radiation layer, which is the multi-standard compatible interface layer of this invention. It contains at least two sets of incompatible heterogeneous ontology standards, namely a proprietary enterprise standard and an internationally accepted standard. The semantic entities of the intermediate layer are mapped to the corresponding ontology classes of different ontology standards through outward connections, so that one intermediate semantic layer can simultaneously interface with multiple sets of heterogeneous ontology standards, reducing the traditional N×M exponential mapping complexity to the linear N+M level.

[0058] Figure 3 Taking cross-standard full-domain query of "loan contract" data as the core scenario, the invention fully demonstrates the execution process, conflict resolution capability and actual technical effect of cross-standard retrieval, which corresponds one by one with the complete steps of cross-standard data retrieval in the claims.

[0059] Figure 3 Using ERP operation applications and compliance and regulatory reports as the two major business scenarios for querying, the semantic entity of the loan contract in the middle layer as the core hub, the physical data table of the underlying loan contract as the unified data source, and the external standard library as the standard adaptation basis, it fully presents the same underlying business data, while adapting to two incompatible ontology standards and meeting the query needs of different business scenarios throughout the entire process.

[0060] Path A represents the internal operations view query process, corresponding to a query scenario based on the enterprise's proprietary standards. When the ERP application initiates a contract data query command based on the enterprise's proprietary standards, the system traces the corresponding subclass semantic entities through the middle layer, routes to the underlying unified physical data table to extract the source data, and finally returns contract data that conforms to the enterprise's proprietary standard format, completing the cross-standard query for the internal operations scenario.

[0061] Path B represents the external regulatory view query process, corresponding to a query scenario based on internationally accepted standards. When a compliance report initiates a query instruction for loan agreement data based on international standards, the system traces the corresponding equivalence class semantic entity through an intermediate layer, routes to the same underlying physical data table to extract source data, and finally returns contract data conforming to the international standard format, completing the cross-standard query for external regulatory scenarios.

[0062] like Figure 4 As shown, an embodiment of this application provides a multi-source heterogeneous ontology fusion device 300, including: a first acquisition module 310, a first determination module 320, a second determination module 330, a third determination module 340, a fourth determination module 350, a fifth determination module 360, a sixth determination module 370, and a seventh determination module 380; the first acquisition module 310 is used to acquire structured metadata of multiple heterogeneous physical data tables, and ontology structured data of at least two sets of heterogeneous ontology standards; the first determination module 320 is used to construct an intermediate semantic layer independent of the heterogeneous physical data tables and heterogeneous ontology standards based on the structured metadata and ontology structured data; the second determination module 330 is used to map each heterogeneous physical data table to the intermediate semantic layer to generate... The system generates a first mapping relation dataset; a third determining module 340 maps the intermediate semantic layer to various heterogeneous ontology standards to generate a second mapping relation dataset; a fourth determining module 350 constructs a unified knowledge graph ontology library based on the first and second mapping relation datasets; a fifth determining module 360 ​​determines the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library in response to cross-ontology standard data retrieval instructions; a sixth determining module 370 routes the data to the corresponding heterogeneous physical data table according to the mapping path and extracts source data from the heterogeneous physical data table; and a seventh determining module 380 performs field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard, and outputs structured retrieval result data.

[0063] The multi-source heterogeneous ontology fusion device 300 acquires structured metadata of heterogeneous physical data tables and ontology structured data of multiple sets of heterogeneous ontology standards through the first acquisition module 310. The first determination module 320 constructs an intermediate semantic layer independent of the physical tables and external standards based on this. The second determination module 330 maps each heterogeneous physical data table to the intermediate semantic layer to generate a first mapping relationship dataset. The third determination module 340 maps the intermediate semantic layer to each heterogeneous ontology standard to generate a second mapping relationship dataset. The fourth determination module 350 constructs a unified knowledge graph ontology library based on the two mapping datasets, thereby transforming the traditional point-to-point multiplicative mapping complexity into... The complexity is reduced to an additive mapping with an intermediate semantic layer as the hub. When the fifth determining module 360 ​​receives a cross-ontology standard data retrieval instruction, it determines the mapping path corresponding to the target standard based on the unified knowledge graph ontology library. The sixth determining module 370 routes to the corresponding physical table to extract source data according to the path. The seventh determining module 380 performs field transformation and relation reorganization according to the target standard and outputs structured results. This realizes plug-and-play extension of any number of heterogeneous ontology standards without modifying the existing physical layer mapping, which significantly reduces the maintenance cost of multi-standard semantic fusion, while ensuring the accuracy and consistency of cross-standard retrieval.

[0064] like Figure 5 As shown, an embodiment of this application provides a multi-source heterogeneous ontology fusion apparatus 400, including a processor 402 and a memory 404. The memory 404 stores a program or instructions. When the processor 402 executes the program or instructions in the memory 404, it implements the steps of the multi-source heterogeneous ontology fusion method as described in any of the above embodiments. Therefore, the multi-source heterogeneous ontology fusion apparatus 400 possesses all the beneficial effects of the multi-source heterogeneous ontology fusion method as described in any of the above embodiments.

[0065] Embodiments of this application provide a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the multi-source heterogeneous ontology fusion method as described in any of the above embodiments. Therefore, the readable storage medium possesses all the beneficial effects of the multi-source heterogeneous ontology fusion method as described in any of the above embodiments.

[0066] In one embodiment, for ontology classes in external ontology standards for which there is currently no corresponding physical data table, the system does not directly ignore them, but instead uses a pre-defined mechanism and an automatic mapping mechanism with the same name to ensure that no physical node is isolated from the semantic network.

[0067] The specific steps include:

[0068] Pre-configured external standard nodes: When constructing the unified knowledge graph ontology library, for newly added external ontology standards (such as FIBO, BFO, etc.), regardless of whether they have a mapping relationship with existing physical data tables, all ontology classes, attributes and inter-class relationships in the standard are pre-loaded into the unified knowledge graph ontology library as candidate semantic nodes.

[0069] Same name matching detection: When a new heterogeneous physical data table is subsequently connected, the system extracts the table name of the physical table, performs normalization processing (removing prefixes and splitting camelCase / underscores), and matches it with the pre-set external standard node name.

[0070] Automatic generation of intermediate semantic entities: If the physical table name and a certain preset external standard node name meet the preset matching threshold (such as being completely identical or having the same root word), the corresponding semantic entity is automatically generated in the intermediate semantic layer. The semantic entity uses ent: as the namespace and establishes an owl:equivalentClass or rdfs:subClassOf association with the external standard node.

[0071] Dynamically establish mapping relationships: At the same time, the system automatically establishes a mapping relationship between the newly generated semantic entity and the physical data table, updates the first mapping relationship dataset, and thus seamlessly integrates the originally detached physical table into the semantic network.

[0072] The above embodiments further enhance the automated construction capabilities of knowledge graphs, ensuring that even without predefined intermediate layer entities, semantic access to physical tables can be achieved through the guidance of external standard nodes, eliminating "data black holes" and achieving 100% full coverage of the graph. This embodiment is particularly suitable for scenarios where large enterprises frequently access new business systems and new industry standards.

[0073] In the claims, description, and accompanying drawings of this invention, the term "plural" refers to two or more. Unless otherwise explicitly defined, the terms "upper," "lower," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and simplifying the descriptive process, and are not intended to indicate or imply that the device or element referred to must have the described specific orientation, or be constructed and operated in a specific orientation. Therefore, these descriptions should not be construed as limiting the invention. The terms "connected," "installed," "fixed," etc., should be interpreted broadly. For example, "connected" can be a fixed connection between multiple objects, a detachable connection between multiple objects, or an integral connection; it can be a direct connection between multiple objects or an indirect connection between multiple objects through an intermediate medium. For those skilled in the art, the specific meaning of the above terms in this invention can be understood based on the specific circumstances described above.

[0074] In the claims, description, and accompanying drawings of this invention, the terms "one embodiment," "some embodiments," "specific embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In the claims, description, and accompanying drawings of this invention, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0075] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for fusing multi-source heterogeneous ontology, characterized in that, include: Acquire structured metadata from multiple heterogeneous physical data tables, as well as ontology structured data from at least two heterogeneous ontology standards; Based on the structured metadata and the ontology structured data, an intermediate semantic layer independent of the heterogeneous physical data table and the heterogeneous ontology standard is constructed. Each heterogeneous physical data table is mapped to the intermediate semantic layer to generate a first mapping relationship dataset; The intermediate semantic layer is mapped to various heterogeneous ontology standards to generate a second mapping relationship dataset; A unified knowledge graph ontology is constructed based on the first mapping relationship dataset and the second mapping relationship dataset; In response to a cross-ontology standard data retrieval command, the mapping path corresponding to the target heterogeneous ontology standard is determined based on the unified knowledge graph ontology library; The mapping path is used to route to the corresponding heterogeneous physical data table, and source data is extracted from the heterogeneous physical data table. The source data is subjected to field transformation and relation reorganization according to the target heterogeneous ontology standard, and structured search result data is output.

2. The method according to claim 1, characterized in that, The structured metadata of the heterogeneous physical data table includes at least one of the following: table name, field name, field type, primary key relationship, foreign key relationship, and inter-table relationship; The ontology structured data of the heterogeneous ontology standard includes at least one of ontology classes, attributes, inter-class relationships, and identification information.

3. The method according to claim 1, characterized in that, The construction of an intermediate semantic layer independent of the heterogeneous physical data table and the heterogeneous ontology standard includes: Configure a unified semantic root node and determine that the semantic classes obtained by mapping each heterogeneous physical data table are all subclasses of the unified semantic root node; At least one basic semantic axis is set under the unified semantic root node, and structured definition data of semantic entities and semantic relations are generated based on the basic semantic axis.

4. The method according to claim 3, characterized in that, The basic semantic axis includes at least one of the following: document semantic axis, economic semantic axis, process semantic axis, subject semantic axis, rule semantic axis, master data semantic axis, and detail semantic axis.

5. The method according to claim 3, characterized in that, Also includes: A composite semantic axis is generated based on at least two basic semantic axes, and the corresponding heterogeneous physical data table is mapped to the basic semantic axis or the composite semantic axis according to the table characteristics and field characteristics of the heterogeneous physical data table.

6. The method according to claim 3, characterized in that, The step of mapping the intermediate semantic layer to various heterogeneous ontology standards to generate a second mapping relationship dataset includes: Based on the matching results between the semantic entities and the ontology classes, attributes, and inter-class relationships in various heterogeneous ontology standards, cross-standard mapping relationships are established according to ontology subclass relationships and / or ontology equivalence class relationships.

7. The method according to claim 1, characterized in that, It also includes a processing step for adding a heterogeneous ontology standard, the processing step including: Obtain the newly added heterogeneous ontology standard file and parse it to obtain the newly added ontology structured data; Configure global priority scores for each heterogeneous ontology standard; When the newly added ontology class conflicts with the identification information and name of the semantic entity in the intermediate semantic layer, the conflict handling result is determined according to the corresponding global priority score. When a new heterogeneous ontology standard is added, the first mapping relationship dataset remains unchanged; Based on the conflict resolution results, the fusion and deduplication are completed, and the unified knowledge graph ontology is updated.

8. A multi-source heterogeneous body fusion device, characterized in that, include: The first acquisition module is used to acquire structured metadata of multiple heterogeneous physical data tables, as well as ontology structured data of at least two heterogeneous ontology standards. The first determining module is used to construct an intermediate semantic layer independent of the heterogeneous physical data table and the heterogeneous ontology standard based on the structured metadata and the ontology structured data. The second determining module is used to map each heterogeneous physical data table to the intermediate semantic layer to generate a first mapping relationship dataset. The third determining module is used to map the intermediate semantic layer to various heterogeneous ontology standards to generate a second mapping relationship dataset. The fourth determining module is used to construct a unified knowledge graph ontology based on the first mapping relationship dataset and the second mapping relationship dataset; The fifth determination module is used to determine the mapping path corresponding to the target heterogeneous ontology standard based on the unified knowledge graph ontology library in response to cross-ontology standard data retrieval instructions. The sixth determining module is used to route to the corresponding heterogeneous physical data table according to the mapping path, and extract source data from the heterogeneous physical data table; The seventh determination module is used to perform field transformation and relation reorganization on the source data according to the target heterogeneous ontology standard, and output structured search result data.

9. A multi-source heterogeneous body fusion device, characterized in that, include: processor; A memory storing programs or instructions, wherein the processor, when executing the programs or instructions in the memory, implements the steps of the multi-source heterogeneous ontology fusion method as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the multi-source heterogeneous ontology fusion method as described in any one of claims 1 to 7.

Images