Method, device and equipment for generating and checking multi-source bloodline candidate link and medium

Abstract

The application discloses a multi-source bloodline candidate link generation and verification method, device, equipment and medium. The method comprises the following steps: S1, uniformly modeling multi-source bloodline fragments, comprehensively calculating attribute similarity, hierarchical mapping consistency and relationship constraint satisfaction degree, and generating a dynamic weighted candidate link set; S2, introducing a constraint breadth-first search algorithm on the basis of the candidate link set, combining a conflict detection and loop verification mechanism to perform link expansion; S3, performing multi-dimensional credibility evaluation on the inferred link, and feeding back the verification result to the candidate link generation link, automatically adjusting the weight parameter and constraint threshold of the candidate link generation stage, thereby forming a complete closed-loop optimization system. The application automatically constructs and verifies multi-source candidate links in the data bloodline knowledge graph, not only improves the integrity and accuracy of link discovery, but also significantly enhances the real-time verification and risk warning capability in the data cross-domain flow scene.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular, to a method, apparatus, device, and medium for generating and verifying multi-source lineage candidate links. Background Technology

[0002] In modern information systems, data link processing faces higher demands. To improve the reliability and integrity of data links, researchers have proposed various algorithmic solutions. However, existing technologies still have shortcomings in monitoring and completing data links. Link matching is coarse, completed links lack constraints, and the reliability of results is insufficient. It is difficult to completely and accurately identify the dependencies of data across multiple systems and levels, leading to breaks or omissions in the data flow links. This affects data traceability and processing quality, and fails to meet the needs for reliable traceability and compliance auditing of data links. Summary of the Invention

[0003] This application provides a method for generating and verifying multi-source lineage candidate links to solve the technical problem that existing technologies are unable to completely and accurately identify the dependencies between multiple systems and multiple levels of data, resulting in the breakage or omission of data flow links, thereby affecting the quality of data traceability and processing.

[0004] This application is achieved through the following solution:

[0005] A method for generating and verifying multi-source kinship candidate links, including the following steps:

[0006] S1. Perform unified modeling on multi-source bloodline fragments, comprehensively calculate attribute similarity, hierarchical mapping consistency and relation constraint satisfaction, and generate a dynamically weighted candidate link set;

[0007] S2. Based on the candidate link set, a constrained breadth-first search algorithm is introduced, combined with conflict detection and loop verification mechanisms, to gradually expand potential links and eliminate redundancy or false dependencies.

[0008] S3. Perform multi-dimensional credibility evaluation on the links obtained from reasoning, including attribute integrity, relationship conformity and hierarchical consistency indicators, eliminate low-confidence links, and simultaneously feed the verification results back to the candidate link generation stage, automatically adjust the weight parameters and constraint thresholds of the candidate link generation stage, thereby forming a complete closed-loop optimization system.

[0009] Further, step S1 specifically includes the following steps:

[0010] S11. Calculate the attribute similarity of nodes across multiple dimensions using a weighted method, thereby quantifying the degree of association between different nodes at the attribute level and achieving cross-system attribute alignment:

[0011] ;

[0012] Where u and v represent candidate nodes; n represents the number of attribute dimensions; w i Indicates the first The weights of each attribute; δ i (u i v i ) indicates the first For each attribute similarity function, Euclidean distance is used for numerical data and cosine similarity is used for textual data; u i v i Indicates the node at the 1st The possible values ​​for each attribute;

[0013] S12. By comparing the hierarchical differences between the nodes, measure the hierarchical consistency of the two in terms of hierarchical structure:

[0014] ;

[0015] Where level(u) and level(v) represent the candidate node level labels; max_level represents the maximum level value in the system;

[0016] S13. By introducing relation constraint functions, the legality of candidate links is verified, and the relation constraint satisfaction degree is calculated to ensure that the generated links conform to predefined business logic or dependency rules:

[0017] ;

[0018] Among them, R pred Φ represents a predefined set of relational rules; r (u, v) represents the relation constraint function, which is 1 if the constraint is satisfied and 0 otherwise.

[0019] S14. Dynamically weight and fuse the three types of information—attribute, hierarchy, and relational constraints—to generate a comprehensive score for candidate links:

[0020] ;

[0021] in, This indicates dynamic weights that adjust with iteration;

[0022] S15. By combining the matching scores of attributes, hierarchy, and relationship constraints, select node pairs that exceed the threshold and generate a candidate link set:

[0023] ;

[0024] Among them, E cand τ represents the set of candidate links; M This represents the threshold for generating candidate links.

[0025] Furthermore, step S2 specifically includes the following steps:

[0026] S21. Based on the constraint-driven graph search method, starting from the preset starting node, and combining attribute, level, and relation constraints, the candidate link graph is traversed step by step to expand the candidate links and generate an expanded set of candidate links:

[0027] ;

[0028] Among them, G cand =(v,E cand BFS Φ G represents a breadth-first search with constraints Φ; cand Represents the candidate link graph; v represents a candidate node; S start Represents the set of starting nodes; E cand Represents the set of candidate links;

[0029] S22. Perform multi-dimensional constraint verification on the expanded candidate link set and eliminate paths that do not meet the constraints:

[0030] ;

[0031] ;

[0032] Where S(u, v), L(u, v), and R(u, v) represent attribute similarity, hierarchical consistency, and relation constraint satisfaction, respectively; τ attr , τ level , τ R These are attribute, level, and relationship threshold, respectively. The indicator function is represented by ψ(u, v); the path constraint filtering function is represented by P. valid This represents the set of valid paths filtered by constraints.

[0033] S23. Detect loops in the network and remove redundant or circular links:

[0034] ;

[0035] Where I represents the identity matrix; A represents the matrix formed by P valid The constructed adjacency matrix; L cycle This represents the determinant value, used to determine whether a loop exists in the network; a non-zero value indicates the existence of a loop.

[0036] S24. Perform conflict analysis on the candidate link set to identify and eliminate links with inconsistent attributes, levels, or relationships:

[0037] ;

[0038] ;

[0039] in, :C confilict (u, v) indicates a conflict, 1 indicates a conflict, and 0 indicates no conflict; P final This represents the set of candidate links that ultimately pass the consistency check.

[0040] Furthermore, step S3 specifically includes the following steps:

[0041] S31. Perform attribute integrity verification on each candidate link that has passed the consistency check, and evaluate the degree of matching of the node attributes involved in each link:

[0042] ;

[0043] Among them, matched attributes represent the number of node attributes that successfully matched; total attributes represent the total number of node attributes.

[0044] S32. Based on attribute integrity verification, by comparing the matching of predefined relationship rules with the actual link relationships, evaluate the relationship compliance between each node in the link and eliminate links that do not conform to business logic or constraints:

[0045] ;

[0046] in, This represents the set of relationships for candidate links, with the relationship constraints output by Module 1. Represents relational constraint functions; Indicates an indicator function;

[0047] S33. Perform a consistency check on the hierarchical relationships between nodes in the link to ensure that the hierarchical differences between nodes are within the allowable range and to avoid unreasonable upstream and downstream relationships:

[0048] ;

[0049] in, Indicates node hierarchy labels; Indicates the maximum level of the system; Indicates the link-level consistency score;

[0050] S34. The verification results of attribute completeness, relationship conformity, and hierarchical consistency are comprehensively calculated to generate a multi-dimensional comprehensive confidence score for each link:

[0051] ;

[0052] in, Represents the weight coefficients of each dimension, satisfying ;

[0053] This represents the multi-dimensional comprehensive confidence score for each link;

[0054] S35. The step of feeding back low-confidence or non-compliant link information to generate a dynamically weighted candidate link set, dynamically adjusting weights and constraints to optimize the link generation strategy:

[0055] ;

[0056] in, This indicates that the remaining links after removing low-confidence links form a closed-loop optimization.

[0057] This application also provides a device for generating and verifying multi-source lineage candidate links, including:

[0058] The candidate link generation module is used to perform unified modeling of multi-source bloodline fragments, comprehensively calculate attribute similarity, hierarchical mapping consistency and relation constraint satisfaction, and form a dynamically weighted candidate link set;

[0059] The constraint-driven graph reasoning module is used to introduce a constraint breadth-first search algorithm on the basis of the candidate link set, combined with conflict detection and loop verification mechanisms, to gradually expand potential links and eliminate redundancy or false dependencies.

[0060] The multi-dimensional link verification and feedback module is used to evaluate the credibility of the inferred links in multiple dimensions, including attribute integrity, relationship conformity and hierarchical consistency indicators. Low-confidence links are eliminated, and the verification results are fed back to the candidate link generation stage. The weight parameters and constraint thresholds of the candidate link generation stage are automatically adjusted to form a complete closed-loop optimization system.

[0061] Furthermore, the candidate link generation module specifically includes:

[0062] The attribute similarity calculation module is used to calculate the attribute similarity of nodes across multiple dimensions using a weighted method, thereby quantifying the degree of association between different nodes at the attribute level and achieving cross-system attribute alignment.

[0063] ;

[0064] Where u and v represent candidate nodes; n represents the number of attribute dimensions; w i Indicates the first The weights of each attribute; δ i (u i v i ) indicates the first For each attribute similarity function, Euclidean distance is used for numerical data and cosine similarity is used for textual data; u i v i Indicates the node at the 1st The possible values ​​for each attribute;

[0065] The hierarchical consistency calculation module measures the hierarchical consistency between nodes by comparing their hierarchical differences.

[0066] ;

[0067] Where level(u) and level(v) represent the candidate node level labels; max_level represents the maximum level value in the system;

[0068] The validity verification module is used to verify the validity of candidate links by introducing relation constraint functions, calculate the relation constraint satisfaction degree, and ensure that the generated links conform to predefined business logic or dependency rules.

[0069] ;

[0070] Among them, R pred Φ represents a predefined set of relational rules; r (u, v) represents the relation constraint function, which is 1 if the constraint is satisfied and 0 otherwise.

[0071] The comprehensive score calculation module dynamically weights and fuses three types of information—attributes, hierarchy, and relational constraints—to generate a comprehensive score for candidate links.

[0072] ;

[0073] in, This indicates dynamic weights that adjust with iteration;

[0074] The candidate link filtering module uses matching scores based on comprehensive attributes, hierarchy, and relational constraints to filter out node pairs that exceed a threshold, generating a candidate link set.

[0075] ;

[0076] Among them, E cand τ represents the set of candidate links; M This represents the threshold for generating candidate links.

[0077] Furthermore, the constraint-driven graph reasoning module specifically includes:

[0078] The candidate link expansion module is used to expand the candidate links by starting from a preset starting node and combining attribute, level, and relation constraints according to a constraint-driven graph search method, and generating an expanded set of candidate links.

[0079] ;

[0080] Among them, G cand =(v,E cand BFS Φ G represents a breadth-first search with constraints Φ; cand Represents the candidate link graph; v represents a candidate node; S start Represents the set of starting nodes; E cand Represents the set of candidate links;

[0081] The path multidimensional constraint verification module is used to perform multidimensional constraint verification on the expanded candidate link set and eliminate paths that do not meet the constraints.

[0082] ;

[0083] ;

[0084] Where S(u, v), L(u, v), and R(u, v) represent attribute similarity, hierarchical consistency, and relation constraint satisfaction, respectively; τ attr , τ level , τ R These are attribute, level, and relationship threshold, respectively. The indicator function is represented by ψ(u, v); the path constraint filtering function is represented by P. valid This represents the set of valid paths filtered by constraints.

[0085] The loop detection module is used to detect loops in the network and remove redundant or circular links.

[0086] ;

[0087] Where I represents the identity matrix; A represents the matrix formed by P valid The constructed adjacency matrix; L cycle This represents the determinant value, used to determine whether a loop exists in the network; a non-zero value indicates the existence of a loop.

[0088] The conflict analysis module is used to perform conflict analysis on the candidate link set, identifying and eliminating links with inconsistent attributes, levels, or relationships.

[0089] ;

[0090] ;

[0091] in, :C confilict (u, v) indicates a conflict, 1 indicates a conflict, and 0 indicates no conflict; P final This represents the set of candidate links that ultimately pass the consistency check.

[0092] Furthermore, the multi-dimensional link verification and feedback module specifically includes:

[0093] The attribute integrity verification module is used to verify the attribute integrity of each candidate link that has passed the consistency verification, and to evaluate the degree of matching of the node attributes involved in each link:

[0094] ;

[0095] Among them, matched attributes represent the number of node attributes that successfully matched; total attributes represent the total number of node attributes.

[0096] The relationship compliance verification module, based on attribute integrity verification, evaluates the relationship compliance between nodes in the link by comparing the matching of predefined relationship rules with the actual link relationships, and eliminates links that do not conform to business logic or constraints.

[0097] ;

[0098] in, This represents the set of relationships for candidate links, with the relationship constraints output by Module 1. Represents relational constraint functions; Indicates an indicator function;

[0099] The hierarchical consistency verification module is used to verify the consistency of the hierarchical relationship between nodes in the link, ensuring that the hierarchical differences between nodes are within the allowable range and avoiding unreasonable upstream and downstream relationships.

[0100] ;

[0101] in, Indicates node hierarchy labels; Indicates the maximum level of the system; Indicates the link-level consistency score;

[0102] The multi-dimensional comprehensive confidence calculation module integrates the verification results of three dimensions—attribute completeness, relationship conformity, and hierarchical consistency—to generate a multi-dimensional comprehensive confidence score for each link.

[0103] ;

[0104] in, Represents the weight coefficients of each dimension, satisfying ;

[0105] This represents the multi-dimensional comprehensive confidence score for each link;

[0106] The feedback and link optimization module is used to feed back low-confidence or non-compliant link information to the step of generating a dynamically weighted candidate link set, dynamically adjusting weights and constraints to optimize the link generation strategy.

[0107] ;

[0108] in, This indicates that the remaining links after removing low-confidence links form a closed-loop optimization.

[0109] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method for generating and verifying multi-source lineage candidate links.

[0110] This application also provides a storage medium including a stored program that, when the program is executed, controls the device where the storage medium is located to perform the steps of the method for generating and verifying multi-source lineage candidate links.

[0111] Compared with the prior art, this application has the following advantages:

[0112] I. Achieving Dynamic Multi-Constraint Matching. This application proposes an innovative dynamic multi-constraint matching mechanism that organically combines attribute similarity, hierarchical mapping consistency, and relation constraint satisfaction, optimizing the priority of candidate links through dynamic weight adjustment. This mechanism can accurately capture heterogeneous features between different systems and automatically balance multi-dimensional constraints, making link generation not only efficient but also flexible. In practical operation, this method can quickly aggregate scattered lineage fragments to generate a high-quality candidate link set, providing a solid foundation for subsequent inference and network expansion, and significantly improving the automation level and accuracy of cross-system lineage analysis;

[0113] II. Global Consistency Maintenance Capability. To ensure the reliability of complex lineage networks in multi-level, multi-relationship environments, this application designs a constraint-driven graph reasoning module. Combined with path constraint filtering, loop detection, and conflict resolution mechanisms, it achieves global consistency maintenance. This module not only effectively avoids redundant links and logical conflicts but also dynamically monitors structural rationality during network expansion. In practical applications, this consistency maintenance mechanism guarantees the integrity and traceability of the lineage network topology, making cross-system data dependencies clear and transparent, and improving the credibility of network analysis and queries.

[0114] Third, it supports multi-dimensional credibility assessment. This application introduces a multi-dimensional link verification mechanism to comprehensively evaluate candidate links from three aspects: attribute integrity, relationship conformity, and hierarchical consistency. The overall credibility of the link is then derived through weighted summation. This design not only effectively eliminates low-confidence links but also feeds the verification results back to the candidate link generation module in real time, forming a dynamic optimization loop. In real-world business scenarios, this mechanism ensures that the lineage network maintains high accuracy and robustness, providing reliable assurance for data tracing and dependency analysis in critical tasks.

[0115] IV. Formation of a Closed-Loop Iterative Optimization System. This application innovatively designs a feedback and link correction mechanism, which eliminates low-confidence links while automatically adjusting the weight parameters and constraint thresholds in the candidate link generation stage, thereby forming a complete closed-loop optimization system. This system can continuously iterate and evolve during network construction, enabling the lineage network to continuously improve in quality while expanding in scale. In actual operation, this iterative optimization significantly improves the system's ability to cope with complex data environments, realizing the self-adaptation and self-reinforcement of lineage network construction;

[0116] V. Achieving Cross-System End-to-End Tracing. Through the organic combination of dynamic constraint matching, graph reasoning consistency maintenance, and multi-dimensional credibility assessment, this application ultimately achieves the complete construction and optimization of a cross-system lineage network. The generated network is not only structurally sound and clearly defined, but also possesses high semantic consistency and topological stability, enabling it to support cross-system end-to-end data tracing and lifecycle management. In complex scenarios such as power and finance that require global dependency analysis, this system can provide a high-quality lineage view, improve business transparency and decision-making intelligence, and has significant practical application value.

[0117] In addition to the purposes, features, and advantages described above, this application has other purposes, features, and advantages. A further detailed description of this application will be provided below with reference to the figures. Attached Figure Description

[0118] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0119] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein:

[0120] Figure 1 This is a flowchart illustrating the method for generating and verifying multi-source lineage candidate links according to a preferred embodiment of this application;

[0121] Figure 2 This is a schematic diagram of the multi-source lineage candidate link generation and verification device module according to a preferred embodiment of this application;

[0122] Figure 3 This is a schematic block diagram of an electronic device according to a preferred embodiment of this application;

[0123] Figure 4 This is an internal structural diagram of a computer device according to a preferred embodiment of this application. Detailed Implementation

[0124] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0125] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0126] It should be noted that the execution subject in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or a device for generating and verifying multi-source lineage candidate links capable of performing the above functions. The following description uses a device for generating and verifying multi-source lineage candidate links as the execution subject to illustrate this embodiment and the subsequent embodiments.

[0127] like Figure 1 As shown, a preferred embodiment of this application provides a method for generating and verifying multi-source lineage candidate links, including the following steps:

[0128] S1. A unified model is performed on multi-source bloodline fragments, and the similarity of attributes, the consistency of hierarchical mapping, and the satisfaction of relational constraints are comprehensively calculated to generate a dynamically weighted candidate link set. This step ensures that nodes between different systems can achieve accurate alignment under multi-dimensional features, laying a solid foundation for subsequent graph reasoning and network expansion.

[0129] S2. Based on the candidate link set, a constraint breadth-first search algorithm is introduced, combined with conflict detection and loop verification mechanisms, to gradually expand potential links and eliminate redundant or false dependencies. This step ensures the continuity and traceability of cross-system and multi-level lineage relationships by dynamically applying global consistency constraints during the reasoning process.

[0130] S3. Perform multi-dimensional credibility evaluation on the inferred links, including attribute integrity, relationship conformity and hierarchical consistency indicators, eliminate low-confidence links, and feed the verification results back to the candidate link generation stage. Automatically adjust the weight parameters and constraint thresholds of the candidate link generation stage to form a complete closed-loop optimization system. This step ensures that the final lineage network has high accuracy, robustness and cross-system consistency.

[0131] The multi-source lineage candidate link generation and verification method in this embodiment first lays the foundation for graph reasoning by generating candidate links, ensuring that cross-system nodes are uniformly aligned under attribute, hierarchical, and relational constraints, and forming high-quality candidate links. Next, based on the candidate link results, the network structure is further expanded to ensure the integrity and traceability of the links. Finally, the generated links are evaluated for credibility and feedback is used for correction to ensure the reliability and robustness of the final lineage network. Through this phased processing, this application effectively solves the problems of coarse link matching, lack of constraints in completed links, and insufficient reliability of results in the prior art, improving the overall construction efficiency of cross-system lineage networks and enhancing the reliability of data lineage network traceability.

[0132] As can be seen, the multi-source lineage candidate link generation and verification method in this embodiment is mainly used to improve the efficiency and accuracy of cross-system lineage network construction, meeting the data traceability and dependency analysis needs in complex scenarios. Simultaneously, considering the problems of insufficient candidate link quality, inconsistent reasoning logic, and lack of global constraints in existing technologies for large-scale lineage reasoning, this embodiment specifically integrates graph structure modeling, semantic similarity calculation, and heuristic search algorithms to automatically construct and verify multi-source candidate links in the data lineage knowledge graph. This not only improves the completeness and accuracy of link discovery but also significantly enhances the real-time verification and risk warning capabilities in cross-domain data flow scenarios. This embodiment can be widely applied to fields requiring trusted traceability and compliance auditing of data links, such as power grid data lineage management, financial data governance, and government data sharing.

[0133] Preferably, step S1 specifically includes the following steps:

[0134] S11. Calculate the attribute similarity of nodes across multiple dimensions using a weighted method, thereby quantifying the degree of association between different nodes at the attribute level and achieving cross-system attribute alignment:

[0135] ;

[0136] Where u and v represent candidate nodes; n represents the number of attribute dimensions; w i Indicates the first The weights of each attribute; δ i (u i v i ) indicates the first For each attribute similarity function, Euclidean distance is used for numerical data and cosine similarity is used for textual data; u i v i Indicates the node at the 1st The possible values ​​for each attribute;

[0137] S12. By comparing the hierarchical differences between the nodes, measure the hierarchical consistency of the two in terms of hierarchical structure:

[0138] ;

[0139] Where level(u) and level(v) represent the candidate node level labels; max_level represents the maximum level value in the system;

[0140] This step balances the differences between different system levels through normalization, ensuring that nodes can reflect hierarchical constraints when aligning across systems, and effectively avoiding the problem of "cross-layer incorrect mapping" during link construction.

[0141] S13. By introducing relation constraint functions, the legality of candidate links is verified, and the relation constraint satisfaction degree is calculated to ensure that the generated links conform to predefined business logic or dependency rules:

[0142] ;

[0143] Among them, R pred Φ represents a predefined set of relational rules; r (u, v) represents the relation constraint function, which is 1 if the constraint is satisfied and 0 otherwise.

[0144] This step, by comprehensively calculating the satisfaction of relational constraints, can filter out node pairs that are only similar in attributes or levels but lack actual business connections, thereby improving the accuracy and practicality of the candidate link set.

[0145] S14. Dynamically weight and fuse the three types of information—attribute, hierarchy, and relational constraints—to generate a comprehensive score for candidate links:

[0146] ;

[0147] in, This indicates dynamic weights that adjust with iteration;

[0148] This step dynamically weights and fuses three types of information: attributes, hierarchy, and relational constraints, to generate a comprehensive score for candidate links. The weight parameters are dynamically adjusted during the iteration process, enabling the model to adapt to the feature distribution of different systems or business scenarios. This mechanism avoids the limitations of static weights and improves the flexibility and robustness of candidate link generation.

[0149] S15. By combining the matching scores of attributes, hierarchy, and relationship constraints, select node pairs that exceed the threshold and generate a candidate link set:

[0150] ;

[0151] Among them, E cand τ represents the set of candidate links; M This represents the threshold for generating candidate links.

[0152] Preferably, step S2 specifically includes the following steps:

[0153] S21. Based on the constraint-driven graph search method, starting from the preset starting node, and combining attribute, level, and relation constraints, the candidate link graph is traversed step by step to expand the candidate links and generate an expanded set of candidate links:

[0154] ;

[0155] Among them, G cand =(v,E cand BFS Φ G represents a breadth-first search with constraints Φ; cand Represents the candidate link graph; v represents a candidate node; S start Represents the set of starting nodes; E cand Represents the set of candidate links;

[0156] This step expands the network using a constraint-driven graph search method. The algorithm starts from a preset starting node and, in combination with attribute, level and relation constraints, traverses the candidate link graph step by step to generate possible link paths, ensuring that the expanded network covers potential dependencies and follows the system's internal constraints.

[0157] S22. Perform multi-dimensional constraint verification on the expanded candidate link set and eliminate paths that do not meet the constraints:

[0158] ;

[0159] ;

[0160] Where S(u, v), L(u, v), and R(u, v) represent attribute similarity, hierarchical consistency, and relation constraint satisfaction, respectively; τ attr , τ level , τ R These are attribute, level, and relationship threshold, respectively. The indicator function is represented by ψ(u, v); the path constraint filtering function is represented by P. valid This represents the set of valid paths filtered by constraints.

[0161] This step involves performing multi-dimensional constraint verification on the expanded candidate link set, eliminating paths that do not meet the constraints, and retaining only high-quality and reliable candidate links to ensure the consistency and rationality of the network structure in cross-system and multi-level environments.

[0162] S23. Detect loops in the network and remove redundant or circular links:

[0163] ;

[0164] Where I represents the identity matrix; A represents the matrix formed by P valid The constructed adjacency matrix; L cycle This represents the determinant value, used to determine whether a loop exists in the network; a non-zero value indicates the existence of a loop.

[0165] S24. Perform conflict analysis on the candidate link set to identify and eliminate links with inconsistent attributes, levels, or relationships:

[0166] ;

[0167] ;

[0168] in, :C confilict (u, v) indicates a conflict, 1 indicates a conflict, and 0 indicates no conflict; P final This represents the set of candidate links that ultimately pass the consistency check.

[0169] This step performs conflict analysis on the candidate path set, identifying links with inconsistent attributes, hierarchies, or relationships. Any detected conflicts are eliminated, and relevant information is fed back to the preceding path generation module to dynamically adjust the weights and constraint strategies of the candidate links. This mechanism ensures that the final candidate links maintain global consistency across system environments and provides highly reliable input for the final network construction.

[0170] Preferably, step S3 specifically includes the following steps:

[0171] S31. Perform attribute integrity verification on each candidate link that has passed the consistency check, and evaluate the degree of matching of the node attributes involved in each link:

[0172] ;

[0173] Among them, matched attributes represent the number of node attributes that successfully matched; total attributes represent the total number of node attributes.

[0174] This step, by checking the integrity of attributes, can eliminate low-quality links caused by missing data or inconsistent attributes, thereby ensuring that the nodes connected by the links are accurate and reliable at both the semantic and data levels.

[0175] S32. Based on attribute integrity verification, by comparing the matching of predefined relationship rules with the actual link relationships, evaluate the relationship compliance between each node in the link and eliminate links that do not conform to business logic or constraints:

[0176] ;

[0177] in, This represents the set of relationships for candidate links, with the relationship constraints output by Module 1. Represents relational constraint functions; Indicates an indicator function;

[0178] This step involves evaluating the relationship compliance between nodes in the link and eliminating links that do not conform to business logic or constraints. This step ensures that each edge in the network is not only reasonable in terms of node attributes, but also conforms to business constraints and system specifications at the relationship level.

[0179] S33. Perform a consistency check on the hierarchical relationships between nodes in the link to ensure that the hierarchical differences between nodes are within the allowable range and to avoid unreasonable upstream and downstream relationships:

[0180] ;

[0181] in, Indicates node hierarchy labels; Indicates the maximum level of the system; Indicates the link-level consistency score;

[0182] This step, through hierarchical consistency verification, can reduce logical anomalies and topological chaos, ensuring that the constructed lineage network conforms to the real business scenario in terms of hierarchical structure, and enhancing the rationality of the global topology;

[0183] S34. The verification results of attribute completeness, relationship conformity, and hierarchical consistency are comprehensively calculated to generate a multi-dimensional comprehensive confidence score for each link:

[0184] ;

[0185] in, Represents the weight coefficients of each dimension, satisfying ; This represents the multi-dimensional comprehensive confidence score for each link;

[0186] This step involves weighted integration of evaluation metrics from different dimensions to identify and eliminate low-confidence links while retaining high-confidence links. This provides a reliable basis for network optimization and final output, ensuring the accuracy and robustness of the overall network.

[0187] S35. The step of feeding back low-confidence or non-compliant link information to generate a dynamically weighted candidate link set, dynamically adjusting weights and constraints to optimize the link generation strategy:

[0188] ;

[0189] in, This indicates that the remaining links after removing low-confidence links form a closed-loop optimization.

[0190] This step, while eliminating low-confidence links, automatically adjusts the weight parameters and constraint thresholds during the candidate link generation stage, thus forming a complete closed-loop optimization system. This system can continuously iterate and evolve during network construction, constantly improving link coverage and overall network quality, enabling the lineage network to continuously improve in quality while scaling. In actual operation, this iterative optimization significantly improves the system's ability to cope with complex data environments, achieving adaptive and self-reinforcing lineage network construction.

[0191] like Figure 2 As shown, another preferred embodiment of this application also provides a device for generating and verifying multi-source lineage candidate links, including:

[0192] The candidate link generation module is used to perform unified modeling of multi-source bloodline fragments, comprehensively calculate attribute similarity, hierarchical mapping consistency and relation constraint satisfaction, and generate a dynamically weighted candidate link set.

[0193] The constraint-driven graph reasoning module is used to introduce a constraint breadth-first search algorithm on the basis of the candidate link set, combined with conflict detection and loop verification mechanisms, to gradually expand potential links and eliminate redundancy or false dependencies.

[0194] The multi-dimensional link verification and feedback module is used to evaluate the credibility of the inferred links in multiple dimensions, including attribute integrity, relationship conformity and hierarchical consistency indicators. Low-confidence links are eliminated, and the verification results are fed back to the candidate link generation stage. The weight parameters and constraint thresholds of the candidate link generation stage are automatically adjusted to form a complete closed-loop optimization system.

[0195] The multi-source lineage candidate link generation and verification device provided in this embodiment adopts the multi-source lineage candidate link generation and verification method in the above embodiments, solving the technical problem that the prior art is unable to completely and accurately identify the dependencies of data between multiple systems and multiple levels, resulting in data flow link breaks or omissions, thereby affecting the quality of data traceability and processing. Compared with the prior art, the beneficial effects of the multi-source lineage candidate link generation and verification device provided in this application are the same as the beneficial effects of the multi-source lineage candidate link generation and verification method provided in the above embodiments, and other technical features in the multi-source lineage candidate link generation and verification device are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0196] Preferably, the candidate link generation module specifically includes:

[0197] The attribute similarity calculation module is used to calculate the attribute similarity of nodes across multiple dimensions using a weighted method, thereby quantifying the degree of association between different nodes at the attribute level and achieving cross-system attribute alignment.

[0198] ;

[0199] Where u and v represent candidate nodes; n represents the number of attribute dimensions; w i Indicates the first The weights of each attribute; δ i (u i v i ) indicates the first For each attribute similarity function, Euclidean distance is used for numerical data and cosine similarity is used for textual data; u i v i Indicates the node at the 1st The possible values ​​for each attribute;

[0200] The hierarchical consistency calculation module measures the hierarchical consistency between nodes by comparing their hierarchical differences.

[0201] ;

[0202] Where level(u) and level(v) represent the candidate node level labels; max_level represents the maximum level value in the system;

[0203] The validity verification module is used to verify the validity of candidate links by introducing relation constraint functions, calculate the relation constraint satisfaction degree, and ensure that the generated links conform to predefined business logic or dependency rules.

[0204] ;

[0205] Among them, R pred Φ represents a predefined set of relational rules; r (u, v) represents the relation constraint function, which is 1 if the constraint is satisfied and 0 otherwise.

[0206] The comprehensive score calculation module dynamically weights and fuses three types of information—attributes, hierarchy, and relational constraints—to generate a comprehensive score for candidate links.

[0207] ;

[0208] in, This indicates dynamic weights that adjust with iteration;

[0209] The candidate link filtering module uses matching scores based on comprehensive attributes, hierarchy, and relational constraints to filter out node pairs that exceed a threshold, generating a candidate link set.

[0210] ;

[0211] Among them, E cand τ represents the set of candidate links; M This represents the threshold for generating candidate links.

[0212] Preferably, the constraint-driven graph reasoning module specifically includes:

[0213] The candidate link expansion module is used to expand the candidate links by starting from a preset starting node and combining attribute, level, and relation constraints according to a constraint-driven graph search method, and generating an expanded set of candidate links.

[0214] ;

[0215] Among them, G cand =(v,E cand BFS Φ G represents a breadth-first search with constraints Φ; cand Represents the candidate link graph; v represents a candidate node; S start Represents the set of starting nodes; E cand Represents the set of candidate links;

[0216] The path multidimensional constraint verification module is used to perform multidimensional constraint verification on the expanded candidate link set and eliminate paths that do not meet the constraints.

[0217] ;

[0218] ;

[0219] Where S(u, v), L(u, v), and R(u, v) represent attribute similarity, hierarchical consistency, and relation constraint satisfaction, respectively; τ attr , τ level , τ R These are attribute, level, and relationship threshold, respectively. The indicator function is represented by ψ(u, v); the path constraint filtering function is represented by P. valid This represents the set of valid paths filtered by constraints.

[0220] The loop detection module is used to detect loops in the network and remove redundant or circular links.

[0221] ;

[0222] Where I represents the identity matrix; A represents the matrix formed by P valid The constructed adjacency matrix; L cycle This represents the determinant value, used to determine whether a loop exists in the network; a non-zero value indicates the existence of a loop.

[0223] The conflict analysis module is used to perform conflict analysis on the candidate link set, identifying and eliminating links with inconsistent attributes, levels, or relationships.

[0224] ;

[0225] ;

[0226] in, :C confilict (u, v) indicates a conflict, 1 indicates a conflict, and 0 indicates no conflict; P final This represents the set of candidate links that ultimately pass the consistency check.

[0227] Preferably, the multi-dimensional link verification and feedback module specifically includes:

[0228] The attribute integrity verification module is used to verify the attribute integrity of each candidate link that has passed the consistency verification, and to evaluate the degree of matching of the node attributes involved in each link:

[0229] ;

[0230] Among them, matched attributes represent the number of node attributes that successfully matched; total attributes represent the total number of node attributes.

[0231] The relationship compliance verification module, based on attribute integrity verification, evaluates the relationship compliance between nodes in the link by comparing the matching of predefined relationship rules with the actual link relationships, and eliminates links that do not conform to business logic or constraints.

[0232] ;

[0233] in, This represents the set of relationships for candidate links, with the relationship constraints output by Module 1. Represents relational constraint functions; Indicates an indicator function;

[0234] The hierarchical consistency verification module is used to verify the consistency of the hierarchical relationship between nodes in the link, ensuring that the hierarchical differences between nodes are within the allowable range and avoiding unreasonable upstream and downstream relationships.

[0235] ;

[0236] in, Indicates node hierarchy labels; Indicates the maximum level of the system; Indicates the link-level consistency score;

[0237] The multi-dimensional comprehensive confidence calculation module integrates the verification results of three dimensions—attribute completeness, relationship conformity, and hierarchical consistency—to generate a multi-dimensional comprehensive confidence score for each link.

[0238] ;

[0239] in, Represents the weight coefficients of each dimension, satisfying ;

[0240] This represents the multi-dimensional comprehensive confidence score for each link;

[0241] The feedback and link optimization module is used to feed back low-confidence or non-compliant link information to the step of generating a dynamically weighted candidate link set, dynamically adjusting weights and constraints to optimize the link generation strategy.

[0242] ;

[0243] in, This indicates that the remaining links after removing low-confidence links form a closed-loop optimization.

[0244] like Figure 3 As shown, a preferred embodiment of this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method for generating and verifying multi-source lineage candidate links in the above embodiments.

[0245] This application provides an electronic device that employs the multi-source lineage candidate link generation and verification method described in the above embodiments. This addresses the technical problem in existing technologies where it is difficult to completely and accurately identify data dependencies across multiple systems and levels, leading to breaks or omissions in data flow links, thus affecting data traceability and processing quality. Compared to existing technologies, the beneficial effects of the electronic device provided in this application are the same as those of the multi-source lineage candidate link generation and verification method provided in the above embodiments, and other technical features of the electronic device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0246] like Figure 4 As shown, a preferred embodiment of this application also provides a computer device, which may be a terminal or a liveness detection server, and its internal structure diagram may be as follows. Figure 4 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with other external computer devices via a network connection. When the computer program is executed by the processor, it implements the steps of the above-described method for generating and verifying multi-source lineage candidate links.

[0247] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0248] The computer device provided in this application employs the multi-source lineage candidate link generation and verification method described in the above embodiments, solving the technical problem that existing technologies struggle to completely and accurately identify data dependencies across multiple systems and levels, leading to breaks or omissions in data flow links, thereby affecting data traceability and processing quality. Compared to existing technologies, the beneficial effects of the computer device provided in this application are the same as those of the multi-source lineage candidate link generation and verification method provided in the above embodiments, and other technical features in the electronic device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0249] A preferred embodiment of this application also provides a storage medium, the storage medium including a stored program, which, when the program is executed, controls the device where the storage medium is located to perform the steps of the multi-source lineage candidate link generation and verification method in the above embodiments.

[0250] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0251] If the functions described in this embodiment are implemented as software functional units and sold or used as independent products, they can be stored in one or more computing device-readable storage media. Based on this understanding, the parts of this application's embodiments that contribute to the prior art or the technical solutions can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computing device (which may be a personal computer, server, mobile computing device, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage media include: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0252] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language C++ and the embedded programming language C.

[0253] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0254] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0255] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0256] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the multi-source lineage candidate link generation and verification method described above.

[0257] The computer program product provided in this application solves the technical problem that existing technologies struggle to completely and accurately identify the dependencies between multiple systems and levels of data, leading to breaks or omissions in the data flow chain, thereby affecting the quality of data traceability and processing. Compared with existing technologies, the beneficial effects of the computer program product provided in this application are the same as those of the multi-source lineage candidate link generation and verification method provided in the above embodiments, and will not be elaborated here.

[0258] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0259] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for generating and verifying multi-source kinship candidate links, characterized in that, Including the following steps: S1. Perform unified modeling on multi-source bloodline fragments, comprehensively calculate attribute similarity, hierarchical mapping consistency, and relation constraint satisfaction, and generate a dynamically weighted candidate link set. The attribute similarity represents the degree of association between different nodes at the attribute level; the hierarchical mapping consistency represents the hierarchical consistency of nodes in the hierarchical structure; and the relation constraint satisfaction represents whether the generated link conforms to predefined business logic or dependency rules. S2. Based on the candidate link set, a constraint breadth-first search algorithm is introduced, combined with a conflict detection and loop verification mechanism, to gradually expand potential links and eliminate redundancy or false dependencies. The conflict detection and loop verification mechanism includes constraint verification and conflict analysis of the candidate link set and determining whether there is a loop in the network. S3. Perform a multi-dimensional credibility assessment on the links obtained through reasoning, including attribute completeness, relationship conformity and hierarchical consistency indicators. The attribute completeness refers to the degree of matching of the node attributes involved in each link. The relationship compliance degree represents the matching status between the predefined relationship rules and the actual link relationship; the hierarchical consistency index represents the hierarchical differences between nodes; low-confidence links are eliminated, and the verification results are simultaneously fed back to the candidate link generation stage, automatically adjusting the weight parameters and constraint thresholds of the candidate link generation stage, thereby forming a complete closed-loop optimization system.

2. The method for generating and verifying multi-source kinship candidate links according to claim 1, characterized in that, Step S1 specifically includes the following steps: S11. Calculate the attribute similarity of nodes across multiple dimensions using a weighted method, thereby quantifying the degree of association between different nodes at the attribute level and achieving cross-system attribute alignment: ； Where u and v represent candidate nodes; n represents the number of attribute dimensions; w i Indicates the first The weights of each attribute; δ i (u i v i ) indicates the first For each attribute similarity function, Euclidean distance is used for numerical data and cosine similarity is used for textual data; u i v i Indicates the node at the 1st The possible values ​​for each attribute; S12. By comparing the hierarchical differences between the nodes, measure the hierarchical consistency of the two in terms of hierarchical structure: ； Where level(u) and level(v) represent the candidate node level labels; max_level represents the maximum level value in the system; S13. By introducing relation constraint functions, the legality of candidate links is verified, and the relation constraint satisfaction degree is calculated to ensure that the generated links conform to predefined business logic or dependency rules: ； Among them, R pred Φ represents a predefined set of relational rules; r (u, v) represents the relation constraint function, which is 1 if the constraint is satisfied and 0 otherwise. S14. Dynamically weight and fuse the three types of information—attribute, hierarchy, and relational constraints—to generate a comprehensive score for candidate links: ； in, This indicates dynamic weights that adjust with iteration; S15. By combining the matching scores of attributes, hierarchy, and relationship constraints, select node pairs that exceed the threshold and generate a candidate link set: ； Among them, E cand τ represents the set of candidate links; M This represents the threshold for generating candidate links.

3. The method for generating and verifying multi-source kinship candidate links according to claim 2, characterized in that, Step S2 specifically includes the following steps: S21. Based on the constraint-driven graph search method, starting from the preset starting node, and combining attribute, level, and relation constraints, the candidate link graph is traversed step by step to expand the candidate links and generate an expanded set of candidate links: ； Among them, G cand =(v,E cand BFS Φ G represents a breadth-first search with constraints Φ; cand Represents the candidate link graph; v represents a candidate node; S start Represents the set of starting nodes; E cand Represents the set of candidate links; S22. Perform multi-dimensional constraint verification on the expanded candidate link set and eliminate paths that do not meet the constraints: ； ； Where S(u, v), L(u, v), and R(u, v) represent attribute similarity, hierarchical consistency, and relation constraint satisfaction, respectively; τ attr , τ level , τ R These are attribute, level, and relationship threshold, respectively. The indicator function is represented by ψ(u, v); the path constraint filtering function is represented by P. valid This represents the set of valid paths filtered by constraints. S23. Detect loops in the network and remove redundant or circular links: ； Where I represents the identity matrix; A represents the matrix formed by P valid The constructed adjacency matrix; L cycle This represents the determinant value, used to determine whether a loop exists in the network; a non-zero value indicates the existence of a loop. S24. Perform conflict analysis on the candidate link set to identify and eliminate links with inconsistent attributes, levels, or relationships: ； ； in, :C confilict (u, v) indicates a conflict, 1 indicates a conflict, and 0 indicates no conflict; P final This represents the set of candidate links that ultimately pass the consistency check.

4. The method for generating and verifying multi-source kinship candidate links according to claim 3, characterized in that, Step S3 specifically includes the following steps: S31. Perform attribute integrity verification on each candidate link that has passed the consistency check, and evaluate the degree of matching of the node attributes involved in each link: ； Among them, matched attributes represent the number of node attributes that successfully matched; total attributes represent the total number of node attributes. S32. Based on attribute integrity verification, by comparing the matching of predefined relationship rules with the actual link relationships, evaluate the relationship compliance between each node in the link and eliminate links that do not conform to business logic or constraints: ； in, This represents the set of relationships for candidate links, with the relationship constraints output by Module 1. Represents relational constraint functions; Indicates an indicator function; S33. Perform a consistency check on the hierarchical relationships between nodes in the link to ensure that the hierarchical differences between nodes are within the allowable range and to avoid unreasonable upstream and downstream relationships: ； in, Indicates node hierarchy labels; Indicates the maximum level of the system; Indicates the link-level consistency score; S34. The verification results of attribute completeness, relationship conformity, and hierarchical consistency are comprehensively calculated to generate a multi-dimensional comprehensive confidence score for each link: ； in, Represents the weight coefficients of each dimension, satisfying ; This represents the multi-dimensional comprehensive confidence score for each link; S35. The step of feeding back low-confidence or non-compliant link information to generate a dynamically weighted candidate link set, dynamically adjusting weights and constraints to optimize the link generation strategy: ； in, This indicates that the remaining links after removing low-confidence links form a closed-loop optimization.

5. A device for generating and verifying multi-source kinship candidate links, characterized in that, include: The candidate link generation module is used to perform unified modeling of multi-source lineage fragments, comprehensively calculate attribute similarity, hierarchical mapping consistency, and relation constraint satisfaction, and generate a dynamically weighted set of candidate links. The attribute similarity represents the degree of association between different nodes at the attribute level; the hierarchical mapping consistency represents the hierarchical consistency of nodes in the hierarchical structure; and the relation constraint satisfaction represents whether the generated links conform to predefined business logic or dependency rules. The constraint-driven graph reasoning module is used to introduce a constraint breadth-first search algorithm on the basis of the candidate link set, combined with a conflict detection and loop verification mechanism, to gradually expand potential links and eliminate redundant or false dependencies. The conflict detection and loop verification mechanism includes constraint verification and conflict analysis of the candidate link set and determining whether there is a loop in the network. The multi-dimensional link verification and feedback module is used to perform multi-dimensional credibility evaluation on the inferred links, including attribute integrity, relationship conformity and hierarchical consistency indicators. The attribute integrity refers to the degree of matching of the node attributes involved in each link. The relationship compliance degree represents the matching status between the predefined relationship rules and the actual link relationship; the hierarchical consistency index represents the hierarchical differences between nodes; low-confidence links are eliminated, and the verification results are simultaneously fed back to the candidate link generation stage, automatically adjusting the weight parameters and constraint thresholds of the candidate link generation stage, thereby forming a complete closed-loop optimization system.

6. The device for generating and verifying multi-source kinship candidate links according to claim 5, characterized in that, The candidate link generation module specifically includes: The attribute similarity calculation module is used to calculate the attribute similarity of nodes across multiple dimensions using a weighted method, thereby quantifying the degree of association between different nodes at the attribute level and achieving cross-system attribute alignment. ； Where u and v represent candidate nodes; n represents the number of attribute dimensions; w i Indicates the first The weights of each attribute; δ i (u i v i ) indicates the first For each attribute similarity function, Euclidean distance is used for numerical data and cosine similarity is used for textual data; u i v i Indicates the node at the 1st The possible values ​​for each attribute; The hierarchical consistency calculation module measures the hierarchical consistency between nodes by comparing their hierarchical differences. ； Where level(u) and level(v) represent the candidate node level labels; max_level represents the maximum level value in the system; The validity verification module is used to verify the validity of candidate links by introducing relation constraint functions, calculate the relation constraint satisfaction degree, and ensure that the generated links conform to predefined business logic or dependency rules. ； Among them, R pred Φ represents a predefined set of relational rules; r (u, v) represents the relation constraint function, which is 1 if the constraint is satisfied and 0 otherwise. The comprehensive score calculation module dynamically weights and fuses three types of information—attributes, hierarchy, and relational constraints—to generate a comprehensive score for candidate links. ； in, This indicates dynamic weights that adjust with iteration; The candidate link filtering module uses matching scores based on comprehensive attributes, hierarchy, and relational constraints to filter out node pairs that exceed a threshold, generating a candidate link set. ； Among them, E cand τ represents the set of candidate links; M This represents the threshold for generating candidate links.

7. The device for generating and verifying multi-source kinship candidate links according to claim 6, characterized in that, The constraint-driven graph reasoning module specifically includes: The candidate link expansion module is used to expand the candidate links by starting from a preset starting node and combining attribute, level, and relation constraints according to a constraint-driven graph search method, and generating an expanded set of candidate links. ； Among them, G cand =(v,E cand BFS Φ G represents a breadth-first search with constraints Φ; cand Represents the candidate link graph; v represents a candidate node; S start E represents the set of starting nodes; cand Represents the set of candidate links; The path multidimensional constraint verification module is used to perform multidimensional constraint verification on the expanded candidate link set and eliminate paths that do not meet the constraints. ； ； Where S(u, v), L(u, v), and R(u, v) represent attribute similarity, hierarchical consistency, and relation constraint satisfaction, respectively; τ attr , τ level , τ R These are attribute, level, and relationship threshold, respectively. The indicator function is represented by ψ(u, v); the path constraint filtering function is represented by P. valid This represents the set of valid paths filtered by constraints. The loop detection module is used to detect loops in the network and remove redundant or circular links. ； Where I represents the identity matrix; A represents the matrix formed by P valid The constructed adjacency matrix; L cycle This represents the determinant value, used to determine whether a loop exists in the network; a non-zero value indicates the existence of a loop. The conflict analysis module is used to perform conflict analysis on the candidate link set, identifying and eliminating links with inconsistent attributes, levels, or relationships. ； ； in, :C confilict (u, v) indicates a conflict, 1 indicates a conflict, and 0 indicates no conflict; P final This represents the set of candidate links that ultimately pass the consistency check.

8. The device for generating and verifying multi-source kinship candidate links according to claim 7, characterized in that, The multi-dimensional link verification and feedback module specifically includes: The attribute integrity verification module is used to verify the attribute integrity of each candidate link that has passed the consistency verification, and to evaluate the degree of matching of the node attributes involved in each link: ； Among them, matched attributes represent the number of node attributes that successfully matched; total attributes represent the total number of node attributes. The relationship compliance verification module, based on attribute integrity verification, evaluates the relationship compliance between nodes in the link by comparing the matching of predefined relationship rules with the actual link relationships, and eliminates links that do not conform to business logic or constraints. ； in, This represents the set of relationships for candidate links, with the relationship constraints output by Module 1. Represents relational constraint functions; Indicates an indicator function; The hierarchical consistency verification module is used to verify the consistency of the hierarchical relationship between nodes in the link, ensuring that the hierarchical differences between nodes are within the allowable range and avoiding unreasonable upstream and downstream relationships. ； in, Indicates node hierarchy labels; Indicates the maximum level of the system; Indicates the link-level consistency score; The multi-dimensional comprehensive confidence calculation module integrates the verification results of three dimensions—attribute completeness, relationship conformity, and hierarchical consistency—to generate a multi-dimensional comprehensive confidence score for each link. ； in, Represents the weight coefficients of each dimension, satisfying ; This represents the multi-dimensional comprehensive confidence score for each link; The feedback and link optimization module is used to feed back low-confidence or non-compliant link information to the step of generating a dynamically weighted candidate link set, dynamically adjusting weights and constraints to optimize the link generation strategy. ； in, This indicates that the remaining links after removing low-confidence links form a closed-loop optimization.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for generating and verifying multi-source lineage candidate links as described in any one of claims 1 to 4.

10. A storage medium comprising a stored program, characterized in that, When the program is running, it controls the device where the storage medium is located to perform the steps of the method for generating and verifying multi-source lineage candidate links as described in any one of claims 1 to 4.

Images