Search

CN-121707590-B - False enterprise information identification and tracing method and system based on structured field propagation chain

CN121707590BCN 121707590 BCN121707590 BCN 121707590BCN-121707590-B

Abstract

The invention provides a false enterprise information identification and tracing method and system based on a structured field propagation chain, which comprise the steps of constructing a field propagation chain map based on the relation among fields in an enterprise database, calculating the change trend of a field value and deviation relative to an industry standard or a history standard in terms of the field value of a target field in the field propagation chain map in different time periods, calculating to obtain a field value evolution score, acquiring a plurality of relevant fields related to a propagation path of the target field based on the field propagation chain map, generating a rule gating judgment result for the target field through a configured structured rule engine, carrying out multi-factor fusion calculation based on the field value evolution score, the field confidence and the rule gating judgment result of the target field, obtaining a false information identification result of the target field, and generating a field-level evidence chain corresponding to the identification result.

Inventors

  • WANG RIHONG

Assignees

  • 上海大智慧财汇数据科技有限公司

Dates

Publication Date
20260512
Application Date
20260210

Claims (10)

  1. 1. A false enterprise information identification and tracing method based on a structured field propagation chain is characterized by comprising the following steps: S1, constructing a field propagation chain map based on the relation among fields in an enterprise database; Step S2, calculating the change trend of the field value and the deviation relative to an industry standard or a historical standard according to the field value of the target field in the field propagation chain map in different time periods, and calculating the field value evolution score based on the calculated change trend and the relative standard deviation; Step S3, acquiring a plurality of related fields related to a propagation path existing in the target field based on the field propagation chain map, and carrying out joint logic judgment on field values corresponding to the plurality of related fields through a configured structured rule engine to generate a rule gating judgment result aiming at the target field; And S4, carrying out multi-factor fusion calculation based on the field value evolution score, the field confidence and the rule gating judgment result of the target field to obtain a false information identification result of the target field, and generating a field-level evidence chain corresponding to the identification result.
  2. 2. The method for identifying and tracing false enterprise information based on structured field propagation chain according to claim 1, wherein said step S1 comprises: S1.1, acquiring an enterprise field set based on a plurality of heterogeneous databases of an enterprise, and converting an original field into a standard field by utilizing a preset field semantic mapping table; s1.2, selecting a trusted field value meeting preset requirements based on time priority and weight aiming at a standard field set; s1.3, carrying out normalization processing on the optimized field value of the standard field after conversion to obtain a normalized field value; S1.4, constructing an enterprise database based on the standard field and the field value after normalization processing; and S1.5, constructing a directional field propagation chain map based on the calculation dependency relationship, the logic reference relationship or the semantic dependency relationship among the fields in the enterprise database.
  3. 3. The method for identifying and tracing false enterprise information based on structured field propagation chains according to claim 2, wherein the nodes of the field propagation chain graph represent standard fields, the edges of the field propagation chain graph represent relationships between fields, the edge weights of the field propagation chain graph represent value conduction sensitivity or derivative relationships between fields; the directional field propagation chain map is used to represent causal/dependency directions and supports truncated setting of path lengths to limit propagation depth.
  4. 4. The method for identifying and tracing false enterprise information based on structured field propagation chain according to claim 1, wherein said step S2 comprises: S2.1, calculating the change trend of the same target field in different time periods based on time sequence analysis aiming at the field value of the same target field in different time periods, and calculating the relative reference deviation of the same target field based on industry/history references; s2.2, generating a field value evolution score by adopting linear weighting based on the calculated change trend and the relative reference deviation; the method comprises the steps of analyzing and calculating the change trend of the same target field in different time periods based on time sequence, wherein the method comprises the step of dynamically calculating the same ratio/ring ratio change trend of the field by using a time window sliding mechanism, and the change trend comprises a change amplitude and a fluctuation rate.
  5. 5. The method for identifying and tracing false enterprise information based on structured field propagation chain according to claim 1, wherein said step S3 comprises: S3.1, acquiring a plurality of relevant fields associated with the propagation paths of the target fields based on the field propagation chain map; S3.2, a structured rule engine configured according to preset expert rules or preset supervision logic; And step S3.3, carrying out joint judgment on field values corresponding to a plurality of related fields by using a configured structured rule engine, wherein the step S3.3 is to identify whether a path connected by a plurality of fields is a risk path or not as a rule gating judgment result of the target field.
  6. 6. A structured field propagation chain-based false enterprise information identification and tracing system, comprising: a module M1, constructing a field propagation chain map based on the relation among the fields in the enterprise database; The module M2 is used for calculating the change trend of the field value and the deviation relative to the industry standard or the history standard according to the field value of the target field in the field propagation chain map in different time periods, and calculating the field value evolution score based on the calculated change trend and the relative standard deviation; A module M3, acquiring a plurality of related fields related to a propagation path existing in the target field based on the field propagation chain map, and carrying out joint logic judgment on field values corresponding to the plurality of related fields through a configured structured rule engine to generate a rule gating judgment result aiming at the target field; And the module M4 is used for carrying out multi-factor fusion calculation based on the field value evolution score, the field confidence and the rule gating judgment result of the target field to obtain a false information identification result of the target field and generating a field-level evidence chain corresponding to the identification result.
  7. 7. The structured field propagation chain based false enterprise information identification and tracing system of claim 6, wherein said module M1 comprises: the module M1.1 is used for acquiring an enterprise field set based on a plurality of heterogeneous databases of an enterprise and converting an original field into a standard field by utilizing a preset field semantic mapping table; A module M1.2, aiming at a standard field set, selecting a trusted field value meeting preset requirements based on time priority and weight; The module M1.3 performs normalization processing on the optimized field value of the standard field after conversion to obtain a normalized field value; The module M1.4 constructs an enterprise database based on the standard field and the field value after normalization processing; And a module M1.5, constructing a directional field propagation chain map based on the calculation dependency relationship, the logic reference relationship or the semantic dependency relationship among the fields in the enterprise database.
  8. 8. The structured field propagation chain based false enterprise information identification and tracing system of claim 7, wherein the nodes of the field propagation chain graph represent standard fields, the edges of the field propagation chain graph represent relationships between fields, the edge weights of the field propagation chain graph represent value conduction sensitivity or derivative relationships between fields; the directional field propagation chain map is used to represent causal/dependency directions and supports truncated setting of path lengths to limit propagation depth.
  9. 9. The structured field propagation chain based false enterprise information identification and tracing system of claim 6, wherein said module M2 comprises: The module M2.1 is used for analyzing and calculating the change trend of the field value of the same target field in different time periods based on time sequence, and calculating the relative reference deviation of the field value of the same target field based on industry/history references; A module M2.2, generating a field value evolution score by adopting linear weighting based on the calculated change trend and the relative reference deviation; the method comprises the steps of analyzing and calculating the change trend of the same target field in different time periods based on time sequence, wherein the method comprises the step of dynamically calculating the same ratio/ring ratio change trend of the field by using a time window sliding mechanism, and the change trend comprises a change amplitude and a fluctuation rate.
  10. 10. The structured field propagation chain based false enterprise information identification and tracing system of claim 6, wherein said module M3 comprises: a module M3.1, acquiring a plurality of relevant fields associated with the propagation path existing in the target field based on the field propagation chain map; A module M3.2 is a structured rule engine configured according to preset expert rules or preset supervision logic; and the module M3.3 is used for carrying out joint judgment on field values corresponding to a plurality of related fields by adopting a configured structured rule engine, so as to identify whether a path connected by the plurality of fields is a risk path or not as a rule gating judgment result of the target field.

Description

False enterprise information identification and tracing method and system based on structured field propagation chain Technical Field The invention relates to the technical field of artificial intelligence and natural language processing, in particular to a false enterprise information identification and tracing method and system based on a structured field propagation chain. Background Along with the rapid development of market economy, enterprise information is used as a core basis for scenes such as market supervision, credit approval, business cooperation and the like, and the authenticity and the reliability of the enterprise information are very important. However, the current enterprise information presents the characteristics of heterogeneous multisource, frequent updating and complex propagation path, and part of enterprises are in favor of improper interests, and false information is submitted by tampering registered capital, false report management data, false equity structure and other modes, so that market order is disturbed, and related parties such as a supervision department, a financial institution and the like are subjected to significant losses. The existing false enterprise information identification technology is dependent on single field verification, manual auditing or simple rule matching, and has the remarkable limitations that on one hand, potential association and propagation rules among fields are difficult to capture, hidden false information of single field compliance but multi-field joint abnormality cannot be effectively identified, on the other hand, dynamic analysis on a field value evolution process is lacking, reasonable data fluctuation and malicious tampering behaviors are difficult to distinguish, and after false information is identified, propagation sources and influence ranges of the false information cannot be accurately traced, so that supervision and responsibility-following work is difficult to advance efficiently. Therefore, a technical scheme capable of realizing accurate identification and tracing based on field association relation and dynamic evolution features is needed to make up for the defects of the prior art. The patent document CN120217163A (application number: 202510585927.4) discloses a false information detection method and device based on a cross-social media propagation network, which comprises the steps of setting a learnable node selector for a network diagram based on news data, introducing a multi-expert model framework, acquiring a key propagation subgraph by each expert model based on the importance weight of a node output by the node selector, aggregating the key propagation subgraph nodes by adopting an aggregation network, calculating the mutual information loss of the key propagation subgraph based on a graph information bottleneck theory, simultaneously introducing a gating network to output an allocation weight for each expert model and aggregate node aggregation results, performing false information detection classification, constructing classification loss, simultaneously introducing comprehensive loss and diversity loss for the multi-expert model framework, and jointly optimizing the node selector, the gating network and the aggregation network by combining the mutual information loss, thereby improving the accuracy and the effectiveness of false information detection. Disclosure of Invention Aiming at the defects in the prior art, the invention aims to provide a false enterprise information identification and tracing method and system based on a structured field propagation chain. The invention provides a false enterprise information identification and tracing method based on a structured field propagation chain, which comprises the following steps: S1, constructing a field propagation chain map based on the relation among fields in an enterprise database; Step S2, calculating the change trend of the field value and the deviation relative to an industry standard or a historical standard according to the field value of the target field in the field propagation chain map in different time periods, and calculating the field value evolution score based on the calculated change trend and the relative standard deviation; Step S3, acquiring a plurality of related fields related to a propagation path existing in the target field based on the field propagation chain map, and carrying out joint logic judgment on field values corresponding to the plurality of related fields through a configured structured rule engine to generate a rule gating judgment result aiming at the target field; And S4, carrying out multi-factor fusion calculation based on the field value evolution score, the field confidence and the rule gating judgment result of the target field to obtain a false information identification result of the target field, and generating a field-level evidence chain corresponding to the identification result. Preferably, the step S1 incl