A tax intelligent operation and maintenance management system based on a digital base

Abstract

The application discloses a tax intelligent operation and maintenance management system based on a digital and intelligent base, and particularly relates to the technical field of tax business operation and maintenance; the system is constructed by collecting tax operation and maintenance related data of a sharing platform, a tax and bank channel and internal multi-business domains, constructing a cross-domain operation and maintenance object correlation graph, and identifying the relationship change of version solidification and adjacent version difference recognition caliber coding parameter data based on the graph; the system compares and matches to identify the drift candidate set of the cross-domain operation and maintenance object correlation edge, performs the presequence constraint and postsequence landing point check along the caliber coding parameter mapping transmission path to form an influence fragment set, and extracts a semantic abnormal fragment set through semantic consistency review of running representation data; finally, the relationship drift source object is determined according to the co-occurrence convergence characteristics of the relationship change item, so that the accurate identification and rapid positioning of the tax business semantic level anomaly are realized.

Description

Technical Field

[0001] This invention relates to the field of tax business operation and maintenance technology, and more specifically, to a tax intelligent operation and maintenance management system based on a digital intelligence foundation. Background Technology

[0002] Existing tax operations and maintenance management methods typically rely on monitoring the technical status of infrastructure and business systems. Data interaction and business collaboration between tax shared platforms, tax-bank channels, and multi-business domain systems are generally monitored using independent system log recording methods, lacking in-depth analysis and unified monitoring of data transmission anomalies or semantic mismatches caused by changes in data mapping across business domains and systems and changes in business rules.

[0003] Existing technologies lack an effective mechanism for sensing and analyzing the drift in data transmission relationships across systems and business domains caused by changes in tax definitions, business codes, and interface parameters in tax operations. This makes it difficult to identify and accurately locate semantic-level anomalies generated during cross-domain business collaborative processing in a timely manner. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a tax intelligent operation and maintenance management method and system based on a digital intelligence foundation to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A tax intelligent operation and maintenance management system based on a digital infrastructure includes: The graph construction module collects tax operation and maintenance related data from the shared platform, tax and banking channels, and internal multiple business domains, and constructs a cross-domain operation and maintenance object association graph according to the business processing order. Version Differentiation Module: Performs version solidification and adjacent version differential on the cross-domain operation and maintenance object association graph, and generates a set of relationship variation items and a set of stable relationship benchmarks for the corresponding caliber coding parameter data; Drift identification module: It compares and pairs the set of changing relation items with the set of stable relation baselines according to the field source, mapping level and transmission direction to identify the drift candidate set of cross-domain operation and maintenance object related edges; Impact verification module: Performs pre-constraint verification and post-landing point verification along the caliber encoding parameter mapping transmission path corresponding to the drift candidate set to form a cross-domain operation and maintenance object impact fragment set; Semantic verification module: Performs semantic consistency verification on the cross-domain operation and maintenance object impact fragment set and the operation representation data, and extracts the semantically abnormal fragment set; Source localization module: Based on the co-occurrence convergence features of each relation change term in the semantic anomaly fragment set, determine the relation drift source object and output the anomaly localization result.

[0006] In a preferred embodiment, tax operation and maintenance related data are collected from the shared platform, tax-banking channels, and internal multi-business domains. A cross-domain operation and maintenance object association graph is constructed according to the business processing order, specifically as follows: Tax calibers, business codes and interface parameters are read from the shared platform, tax-bank channels and internal multi-business domains and merged into caliber code parameter data. Link events, business result records, rule hit records and operation alarm records are read and merged into operation characterization data. Following the business processing sequence of business acceptance, rule processing, channel interaction, and result feedback, the caliber coding parameter data and operational characterization data are associated with the corresponding cross-domain operation and maintenance objects to generate a cross-domain operation and maintenance object association map.

[0007] In a preferred embodiment, version solidification and adjacent version differentiation are performed on the cross-domain operation and maintenance object association graph to generate a set of relationship variation items and a set of stable relationship benchmarks for the corresponding caliber coding parameter data. Specifically: The version identifier is the formation time and business processing order of the cross-domain operation and maintenance object association graph. The version is then fixed for the cross-domain operation and maintenance objects, caliber coding parameter data and associated edges in the cross-domain operation and maintenance object association graph. The associated edges of corresponding caliber coding parameter data in adjacent solidified versions are differentiated according to the source object, target object, and associated content; Add or remove related edges or change their content and categorize them into the relation change itemset. Consistent related edges are included in the stable relation benchmark set.

[0008] In a preferred embodiment, the set of variable relation itemsets is compared and paired with the set of stable relation baselines according to field source, mapping level, and transmission direction to identify the drift candidate set of cross-domain operation and maintenance object association edges, specifically: Extract the associated edges that have been added, removed, or have changed content from the relational change itemset; Extract stable relation edges from the stable relation benchmark set that have the same source object or the same target object as the associated edges; The associated edges are compared and paired with stable associated edges according to the field source, mapping level, and transmission direction. Associated edges with the same field source but different mapping levels or transmission directions are included in the drift candidate set of cross-domain operation and maintenance object associated edges.

[0009] In a preferred embodiment, preceding constraint verification and subsequent landing point verification are performed along the caliber encoding parameter mapping transmission path corresponding to the drift candidate set to form a cross-domain operation and maintenance object impact fragment set, specifically: Read the cross-domain operation and maintenance object association edges from the drift candidate set, and extract the caliber coding parameter mapping transmission path containing caliber coding parameter data in the cross-domain operation and maintenance object association graph, using the source object and target object of the cross-domain operation and maintenance object association edge as the delimitation endpoints. Perform preceding constraint verification on the preceding cross-domain operation and maintenance object and the stable associated edge of the caliber coding parameter mapping and transmission path. Perform subsequent endpoint verification on the cross-domain operation and maintenance objects and relational change itemsets of the caliber coding parameter mapping and transmission path. Path segments that match both the preceding constraint check and the subsequent landing point check are categorized into the cross-domain operation and maintenance object impact segment set.

[0010] In a preferred embodiment, the semantic consistency of the cross-domain operation and maintenance object impact fragment set and the operational characterization data is reviewed, and the semantically abnormal fragment set is extracted, specifically as follows: Read path segments from the cross-domain operation and maintenance object impact fragment set, and match link events, business result records, rule hit records and operation alarm records in the operation representation data according to the cross-domain operation and maintenance objects and business processing order in the path segments; The matched operational representation data and the caliber coding parameter data in the path segment are semantically consistent in terms of field meaning, business coding affiliation, and interface parameter values. Path segments with inconsistent verification are classified into the semantic anomaly fragment set.

[0011] The technical effects and advantages of the tax intelligent operation and maintenance management system based on digital intelligence foundation of the present invention are as follows: By collecting tax operation and maintenance related data from shared platforms, tax-bank channels, and internal multi-business domains, and constructing a cross-domain operation and maintenance object relationship graph according to the business processing sequence, a traceable relationship foundation is formed between tax caliber, business codes, interface parameters, and operational representation data. By version solidification and adjacent version differentiation of the cross-domain operation and maintenance object relationship graph, stable relationships and relationship changes can be distinguished. By comparing and matching field sources, mapping levels, and transmission directions, the drift candidate set of cross-domain operation and maintenance object relationship edges can be identified. By verifying pre-order constraints and post-order landing points, the path range that relationship drift may affect can be limited. By verifying semantic consistency, semantic anomaly fragments related to relationship drift can be extracted from operational representation data. By determining the source object of relationship drift through co-occurrence convergence features, the accuracy of identifying and the completeness of locating semantic-level anomalies in cross-domain tax business can be improved. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the structure of a tax intelligent operation and maintenance management system based on a digital intelligence foundation according to the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0014] Example 1 Figure 1 This invention presents a tax intelligent operation and maintenance management system based on a digital intelligence foundation, comprising: The graph construction module collects tax operation and maintenance related data from the shared platform, tax and banking channels, and internal multiple business domains, and constructs a cross-domain operation and maintenance object association graph according to the business processing order. Version Differentiation Module: Performs version solidification and adjacent version differential on the cross-domain operation and maintenance object association graph, and generates a set of relationship variation items and a set of stable relationship benchmarks for the corresponding caliber coding parameter data; Drift identification module: It compares and pairs the set of changing relation items with the set of stable relation baselines according to the field source, mapping level and transmission direction to identify the drift candidate set of cross-domain operation and maintenance object related edges; Impact verification module: Performs pre-constraint verification and post-landing point verification along the caliber encoding parameter mapping transmission path corresponding to the drift candidate set to form a cross-domain operation and maintenance object impact fragment set; Semantic verification module: Performs semantic consistency verification on the cross-domain operation and maintenance object impact fragment set and the operation representation data, and extracts the semantically abnormal fragment set; Source localization module: Based on the co-occurrence convergence features of each relation change term in the semantic anomaly fragment set, determine the relation drift source object and output the anomaly localization result.

[0015] Collect tax operation and maintenance related data from the shared platform, tax and banking channels, and internal multiple business domains, and construct a cross-domain operation and maintenance object relationship graph according to the business processing order, including: The scope of tax-related documentation includes definitions of VAT tax rate brackets, mapping relationships of corporate income tax deductions, rules for individual income tax item classification, and applicable conditions for export tax refunds, encompassing related text and structured entries. For tax-bank channels, the documentation also includes field descriptions and message parsing specifications for tax declaration data in bank-tax interaction scenarios. Internal multi-business domain documentation includes localized descriptions of tax processing rules and field meaning annotations for each business domain in the current version. When reading tax-related documentation, the business domain identifier, version number, effective timestamp, and field granularity level of the documentation are recorded. The tax-related documentation, along with the business domain identifier, version number, effective timestamp, and field granularity level, is treated as a subset of the overall stored documentation encoding parameter data.

[0016] The scope of business codes includes tax type codes, tax item sub-item codes, taxpayer category codes, tax incentive policy codes, and tax matter classification codes. Business codes retrieved from the shared platform are stored in a standard code table format, organized in a hierarchical tree structure. The depth of the hierarchy depends on the actual coding system, for example, 3 to 5 levels. Business codes retrieved from tax-bank channels are stored in a message field mapping table format. There is a name conversion relationship between field names and the shared platform's standard codes; this name conversion relationship is recorded synchronously during retrieval. Business codes retrieved from internal multi-business domains are stored in each domain's local code table format. There is a local mapping relationship between local codes and the shared platform's standard codes; this local mapping relationship is recorded synchronously during retrieval. The business codes, name conversion relationships, and local mapping relationships, along with the corresponding business domain identifier and version number, form a subset of the overall stored entry-level coding parameter data.

[0017] The scope of interface parameter reading covers the parameter names, data types, value ranges, required field identifiers, and default value definitions of the interfaces exposed by each business domain. Interface parameters read from the shared platform correspond to shared service interfaces, those read from the tax and banking channels correspond to tax and banking message interfaces, and those read from internal multi-business domains correspond to internal service interfaces within each domain. When reading interface parameters, the interface identifier, version number, and call direction of the interface to which the parameter belongs are recorded. The call direction includes both input and output parameter directions. The interface parameters, along with the interface identifier, version number, and call direction, are stored as a subset of the overall caliber coding parameter data. After independently storing the tax caliber subset, business coding subset, and interface parameter subset, a merging operation is performed on the three subsets using the business domain identifier as the primary key and the version number and effective timestamp as secondary indexes to form the caliber coding parameter data. Each record in the caliber coding parameter data carries the source domain identifier, data category identifier, and data collection timestamp.

[0018] The scope of readable link events covers business request initiation events, service call events, cross-domain data transmission events, and business request termination events. Each link event record includes a link identifier, event occurrence timestamp, event type code, initiating business domain identifier, and receiving business domain identifier. The scope of readable business result records covers tax declaration processing results, tax refund review results, tax-bank interaction receipt results, and business exception return results. Each business result record includes a business serial number, result type code, result generation timestamp, result belonging business domain identifier, and a set of result content fields. The scope of readable rule hit records covers rule hit records generated by the tax rule engine during business request processing. Each rule hit record includes a rule identifier, rule version number, hit timestamp, the hit business domain identifier, and the data version number of the caliber encoding parameter used during the hit. The scope of readable operational alarm records covers abnormal alarm records generated by the shared platform, tax-bank channels, and multiple internal business domains during business processing. Each operational alarm record includes an alarm identifier, alarm type code, alarm occurrence timestamp, alarm belonging business domain identifier, and the link identifier associated with the alarm. After independently reading the four types of data, the link identifier is used as the primary key, and the business serial number, rule identifier, and alarm identifier are used as auxiliary association keys. The link event, business result record, rule hit record, and operation alarm record are merged to form operation characterization data. Each record in the operation characterization data carries the source domain identifier, data category identifier, and data collection timestamp.

[0019] After merging the tax caliber coding parameter data and operational characterization data, the data is associated with the corresponding cross-domain operation and maintenance objects according to the business processing sequence of business acceptance, rule processing, channel interaction, and result feedback, generating a cross-domain operation and maintenance object association graph. Specifically, the cross-domain operation and maintenance object corresponding to the business acceptance stage is the business acceptance object. The business acceptance object undertakes operations such as receiving business requests, verifying taxpayer identities, and preprocessing declaration data. The tax caliber subset records and interface parameter subset records directly associated with the business acceptance stage in the tax caliber coding parameter data, as well as the link event records in the operational characterization data whose event type codes belong to the business request initiation event, are associated with the business acceptance object. The association relationship is represented in the form of directed edges. The source end of the directed edge is the tax caliber coding parameter data record or the operational characterization data record, and the target end is the business acceptance object. The edge attributes of the directed edge include data category identifier, association basis description, and association timestamp.

[0020] The cross-domain operation and maintenance object corresponding to the rule processing stage is the rule processing object. The rule processing object handles operations such as tax rule matching, tax scope applicability judgment, and business code conversion. It associates the tax scope subset records and business code subset records directly related to the rule processing stage in the scope coding parameter data, as well as all rule hit records in the operational representation data, with the association represented by directed edges. The edge attributes of the directed edges include data category identifier, association basis description, and association timestamp. The cross-domain operation and maintenance object corresponding to the channel interaction stage is the channel interaction object. The channel interaction object handles operations such as tax-bank message assembly, tax-bank channel message sending, and tax-bank channel receipt reception. It associates the business code subset records and interface parameter subset records directly related to the channel interaction stage in the scope coding parameter data, as well as the link event records in the operational representation data whose event type code belongs to cross-domain data transmission events and the business result records whose result type code belongs to tax-bank interaction receipt results, with the association represented by directed edges. The edge attributes of the directed edges include data category identifier, association basis description, and association timestamp. The cross-domain operation and maintenance object corresponding to the result feedback stage is the result feedback object. The result feedback object carries out operations such as tax processing result output, result notification issuance, and business closure confirmation. It associates the subset of interface parameters directly related to the result feedback stage in the caliber coding parameter data, as well as the business result records and all operation alarm records in the operation characterization data whose result type codes belong to tax declaration processing results, tax refund review results, or business anomaly return results. The association relationship is represented in the form of directed edges. The edge attributes of the directed edges include data category identifier, association basis description, and association timestamp.

[0021] Following the business processing sequence of business acceptance, rule processing, channel interaction, and result feedback, directed edges are added between the business acceptance object and the rule processing object, between the rule processing object and the channel interaction object, and between the channel interaction object and the result feedback object. The source end of the directed edge is the cross-domain operation and maintenance object that is earlier in the business processing sequence, and the target end is the cross-domain operation and maintenance object that is later in the business processing sequence. The edge attributes of the directed edge include the sequence number and the business processing stage name. The cross-domain operation and maintenance object association graph is organized in the form of a directed graph. The nodes are the business acceptance object, rule processing object, channel interaction object, and result feedback object. The edges are the directed edges between the caliber coding parameter data record and each cross-domain operation and maintenance object, the directed edges between the operation characterization data record and each cross-domain operation and maintenance object, the directed edges between the business processing order of the business acceptance object and the rule processing object, the directed edges between the rule processing object and the channel interaction object, and the directed edges between the channel interaction object and the result feedback object. Each node in the cross-domain operation and maintenance object association graph carries the cross-domain operation and maintenance object identifier, the name of the business processing stage to which the cross-domain operation and maintenance object belongs, and the cross-domain operation and maintenance object creation timestamp. Each edge carries the edge type identifier, the edge attribute set, and the edge creation timestamp. The cross-domain operation and maintenance object association graph is persistently stored in the form of a graph database, using the cross-domain operation and maintenance object identifier and the edge type identifier as the query index.

[0022] Version solidification and adjacent version differentiation are performed on the cross-domain operation and maintenance object relationship graph to generate a set of relationship variation items and a set of stable relationship benchmarks for the corresponding caliber coding parameter data, including: The version identifier is composed of the formation time of the cross-domain operation and maintenance object association graph and the business processing sequence. The formation time is the system timestamp when the cross-domain operation and maintenance object association graph completes the persistent storage operation, ensuring that the cross-domain operation and maintenance object association graphs triggered by different batches of collection are uniquely distinguishable in the time dimension. The business processing sequence is the sequence number corresponding to the four defined business processing stages: business acceptance, rule processing, channel interaction, and result feedback. The sequence number is stored in the cross-domain operation and maintenance object association graph as the edge attribute of the directed edge of the business processing sequence. The version identifier is expressed as a combined string of the formation timestamp and the business processing sequence number, for example, in the concatenation format of "formation timestamp_sequence number". The version identifier is written as an independent field in the graph database storage record of the cross-domain operation and maintenance object association graph. A unique index is established in the graph database for the version identifier field to ensure that the version identifier of each fixed version is globally unique.

[0023] When performing version fixation on the cross-domain operation and maintenance object association graph, all nodes, edges, and edge attribute sets in the cross-domain operation and maintenance object association graph, along with the version identifier, are completely copied to an independent version storage area in the form of a snapshot. Each fixed version in the version storage area is isolated from each other, and once a fixed version is written, it cannot be overwritten or modified. The trigger condition for version fixation is: after each persistent storage operation of the cross-domain operation and maintenance object association graph is completed, version fixation is immediately performed on the cross-domain operation and maintenance object association graph formed in that batch, ensuring that the fixed version corresponds one-to-one with the batch in which the cross-domain operation and maintenance object association graph was formed. After the version fixation is completed, the version identifier, fixation timestamp, number of cross-domain operation and maintenance objects, number of calibrated parameter data records, and number of associated edges contained in the fixed version are recorded in the version storage area, constituting the version fixation metadata, which is used to determine the two fixed versions participating in the differential analysis when adjacent versions are used.

[0024] The method for determining adjacent fixed versions is as follows: All fixed versions are sorted in ascending order of their formation timestamps from the version storage area. Two fixed versions with adjacent timestamps are taken as the version pair for the differential operation. The fixed version with the smaller timestamp is designated as the preceding fixed version, and the fixed version with the larger timestamp is designated as the following fixed version. If multiple fixed versions with different sequential numbers exist under the same formation timestamp, they are further sorted in ascending order of their business processing sequence numbers within the same formation timestamp, and two fixed versions with adjacent business processing sequence numbers are taken as the version pair. If cross-batch differential operations are performed between different formation timestamps, the fixed version with the largest business processing sequence number among the preceding fixed versions and the fixed version with the smallest business processing sequence number among the following fixed versions are taken as the version pair.

[0025] When performing differential operations on related edges of corresponding caliber coding parameter data in adjacent fixed versions, the processing objects of the differential operation are limited to directed edges whose source or target ends involve caliber coding parameter data records. Specifically, this includes directed edges between caliber coding parameter data records and business acceptance objects, rule processing objects, channel interaction objects, or result feedback objects, as well as directed edges where there are field reference relationships between the caliber coding parameter data records themselves. The differential operation compares the source object, target object, and related content sequentially. The comparison of the source object is based on the combination of the cross-domain operation and maintenance object identifier or the data category identifier of the caliber coding parameter data record carried by the source node of the directed edge and the source domain identifier. The comparison of the target object is based on the combination of the cross-domain operation and maintenance object identifier or the data category identifier of the caliber coding parameter data record carried by the target node of the directed edge and the source domain identifier. The comparison of the related content is based on the edge attribute set of the directed edge, which includes the data category identifier, the description of the association basis, the association timestamp, the edge type identifier, and the edge creation timestamp.

[0026] Between the preceding and following fixed versions, the associated edges of the caliber coding parameter data are paired one by one according to the combination of the source object and the target object as the matching key. If an associated edge exists in the preceding fixed version but not in the following fixed version, it is determined that the associated edge has decreased; if an associated edge does not exist in the preceding fixed version but exists in the following fixed version, it is determined that the associated edge has increased; if an associated edge exists in both the preceding and following fixed versions and the matching key of the source object and the target object is consistent, then the three fields of data category identifier, association basis description, and edge type identifier in the edge attribute set are compared field by field. If the value of any field is inconsistent, it is determined that the associated edge has changed in content. All associated edges determined to have increased, decreased, or changed in content are included in the relation change itemset. Each record in the relation change itemset contains a change type identifier, a complete snapshot of the corresponding associated edge in the preceding fixed version, a complete snapshot of the corresponding associated edge in the following fixed version, a timestamp of the differential operation execution, and a pair of version identifiers participating in the differential operation. The change type identifier can be one of three values: increase, decrease, or content change. In the case of an increase in the complete snapshot of the corresponding associated edge in the previous fixed version, the value is null. In the case of a decrease in the complete snapshot of the corresponding associated edge in the subsequent fixed version, the value is null.

[0027] Edges that exist in both the preceding and subsequent fixed versions, have the same matching key for the source and target objects, and whose data category identifier, association basis description, and edge type identifier values ​​in the edge attribute set are all identical, are considered consistent edges and are included in the stable relation benchmark set. Each record in the stable relation benchmark set contains the source object identifier, target object identifier, edge attribute set, operation timestamp when included in the stable relation benchmark set, and version identifier pair participating in the difference analysis for consistent edges. The stable relation benchmark set uses the combination of the source object identifier and the target object identifier as the retrieval key and establishes a corresponding index in the version storage area of ​​the graph database to support efficient retrieval when comparing and matching relation change itemsets with the stable relation benchmark set according to field source, mapping level, and transit direction.

[0028] The set of variable itemsets is compared and matched with the set of stable relation baselines according to field source, mapping level, and transit direction to identify the drift candidate set of cross-domain operation and maintenance object association edges, including: When extracting associated edges that have been added, removed, or have changed content from the relation change itemset, all records in the relation change itemset are traversed. The records are then filtered according to the value of the change type identifier. All records whose change type identifier is "increase," "decrease," or "change" are selected. For each record, a complete snapshot of the corresponding associated edge in the subsequent fixed version is extracted as the representative form of the associated edge that has been added, removed, or has changed content in the current processing. For records whose change type identifier is "decrease," the complete snapshot of the corresponding associated edge in the subsequent fixed version is null. In this case, the complete snapshot of the corresponding associated edge in the previous fixed version is taken as the representative form of the associated edge that has been added, removed, or has changed content in the current processing. After extraction, each associated edge that has been added, removed, or whose content has changed carries a source object identifier, a target object identifier, an edge attribute set, and a change type identifier. The edge attribute set includes a data category identifier, an association basis description, an association timestamp, an edge type identifier, and an edge creation timestamp. The value format of the source object identifier and the target object identifier is consistent with the matching key defined by the difference operation, that is, the value is a combination of the cross-domain operation and maintenance object identifier or the data category identifier and the source domain identifier carried by the cross-domain operation and maintenance object identifier or the caliber coding parameter data record carried by the directed edge source node.

[0029] When extracting stable relational edges from the stable relation benchmark set that share the same source object or the same target object as the relational edges that have undergone additions, deletions, or content changes, a search operation is performed in the stable relation benchmark set for each relational edge that has undergone additions, deletions, or content changes, using the source object identifier and target object identifier of the relational edge that has undergone additions, deletions, or content changes as search keys. The search scope includes records in the stable relation benchmark set whose source object identifier is the same as the source object identifier of the relational edge that has undergone additions, deletions, or content changes, and records in the stable relation benchmark set whose target object identifier is the same as the target object identifier of the relational edge that has undergone additions, deletions, or content changes. The union of the two types of search results is then used to remove duplicate records, and this union is used as the set of stable relational edges corresponding to the relational edge that has undergone additions, deletions, or content changes. The search operation is performed using the index established in step S2 in the graph database version storage area, which uses the combination of source object identifier and target object identifier as the search key, to ensure search efficiency. Each stable relational edge in the set of stable relational edges carries a source object identifier, a target object identifier, a set of edge attributes, an operation timestamp when it was included in the stable relation benchmark set, and a version identifier pair participating in the difference analysis.

[0030] When comparing and pairing related edges that have been added, removed, or have changed content with stable related edges according to field source, mapping level, and transmission direction, the extraction methods for the three comparison dimensions of field source, mapping level, and transmission direction are first determined. Field source refers to the original source domain identifier of the caliber coding parameter field carried by the related edge in the caliber coding parameter data. The field source is directly read from the source domain identifier carried by the caliber coding parameter data record at the source port of the related edge. When the source end of the related edge is a cross-domain operation and maintenance object rather than a caliber coding parameter data record, the field source is parsed from the association basis description field in the edge attribute set of the related edge. The association basis description field records the original source domain identifier of the caliber coding parameter field in structured text form. The mapping level refers to the depth of the hierarchy that the caliber coding parameter data undergoes after multiple transformations from the original source domain to the current associated target node. The mapping level is obtained by tracing back along the associated edges of the caliber coding parameter data in the cross-domain operation and maintenance object association graph from the current associated target node to the original source domain node corresponding to the field source, and counting the number of intermediate cross-domain operation and maintenance objects passed through in the tracing path plus 1. For example, if the caliber coding parameter data is directly associated from the sharing platform to the rule processing object, the mapping level is 1; if the caliber coding parameter data is first associated from the sharing platform to the business acceptance object, and then associated from the business acceptance object to the rule processing object, the mapping level is 2. The transmission direction refers to the flow of caliber coding parameter data in the business processing sequence dimension of the cross-domain operation and maintenance object association graph. The transmission direction is read from the edge type identifier field in the edge attribute set of the associated edge. The edge type identifier distinguishes between two values: forward transmission and reverse transmission. Forward transmission means that the caliber coding parameter data flows in the business processing sequence direction of business acceptance, rule processing, channel interaction, and result feedback. Reverse transmission means that the caliber coding parameter data flows in the opposite direction of the business processing sequence.

[0031] After extracting the three dimensions of field source, mapping level, and transmission direction, for each added, deleted, or content-changed associated edge, a comparison and pairing is performed sequentially with each stable associated edge in the corresponding stable associated edge set, following the order of field source, mapping level, and transmission direction. The execution logic of the comparison and pairing is as follows: First, compare whether the field source of the added, deleted, or content-changed associated edge is consistent with the field source of the stable associated edge. The criterion for consistent field source is that the source domain identifier values ​​of the two are exactly the same. If the field source is inconsistent, the current comparison and pairing is terminated, and the associated edge that has been added, deleted, or content-changed is not included in the drift candidate set of cross-domain operation and maintenance object associated edges. If the field source is consistent, further compare whether the mapping level of the added, deleted, or content-changed associated edge is consistent with the mapping level of the stable associated edge, and whether the transmission direction of the added, deleted, or content-changed associated edge is consistent with the transmission direction of the stable associated edge. The criterion for consistent mapping level is that the mapping level values ​​of the two are exactly the same, and the criterion for consistent transmission direction is that the edge type identifier values ​​of the two are exactly the same.

[0032] If the fields have the same source but different mapping levels, it is determined that there is a mapping level drift between the added / removed or content-changed associated edges and the stable associated edges, and the associated edges with added / removed or content-changed are included in the drift candidate set of cross-domain operation and maintenance object associated edges. If the fields have the same source but different transmission directions, it is determined that there is a transmission direction drift between the added / removed or content-changed associated edges and the stable associated edges, and the associated edges with added / removed or content-changed are included in the drift candidate set of cross-domain operation and maintenance object associated edges. If the fields have the same source, different mapping levels, and different transmission directions, the associated edges with added / removed or content-changed are also included in the drift candidate set of cross-domain operation and maintenance object associated edges, and both mapping level drift and transmission direction drift are marked in the drift candidate set record. If the fields have the same source, the same mapping level, and the same transmission direction, it is determined that there is no drift between the added / removed or content-changed associated edges and the stable associated edges, and the associated edges with added / removed or content-changed are not included in the drift candidate set of cross-domain operation and maintenance object associated edges. If there are multiple stable associated edges in the set of stable associated edges corresponding to associated edges that have been added, removed, or have changed content, any stable associated edge in the set of stable associated edges that meets the conditions of consistent field source and inconsistent mapping level or transmission direction will be included in the drift candidate set of cross-domain operation and maintenance object associated edges. At the same time, the source object identifier, target object identifier, and edge attribute set of the stable associated edge that triggered the inclusion operation will be recorded in the corresponding record of the drift candidate set of cross-domain operation and maintenance object associated edges as the drift comparison basis.

[0033] Each record in the cross-domain operation and maintenance object's associated edge drift candidate set contains the source object identifier, target object identifier, edge attribute set, change type identifier, drift type identifier, source object identifier of the stable associated edge that triggered the incorporation operation, target object identifier of the stable associated edge that triggered the incorporation operation, edge attribute set of the stable associated edge that triggered the incorporation operation, and the execution timestamp and version identifier pair of the matching operation. The drift type identifier can be one of three: mapping level drift, transmission direction drift, or a combination of mapping level drift and transmission direction drift. The cross-domain operation and maintenance object's associated edge drift candidate set uses the combination of the source object identifier and target object identifier of the associated edge that triggered the incorporation operation or changed content as the retrieval key, and establishes a corresponding index in the version storage area of ​​the graph database to support efficient read operations when performing pre-constraint verification and post-landing point verification along the mapping transmission path corresponding to the drift candidate set's caliber encoding parameters. After the drift candidate set of the cross-domain operation and maintenance object association edge is written, the total number of drift candidate set records and the version identifier pair corresponding to this comparison and matching operation are recorded and stored independently as drift candidate set metadata.

[0034] Along the caliber encoding parameter mapping transmission path corresponding to the drift candidate set, perform pre-constraint verification and post-landing point verification to form a cross-domain operation and maintenance object impact fragment set, including: When reading cross-domain operation and maintenance object (O&M) related edges from the drift candidate set, all records in the drift candidate set are traversed. From each record, the source object identifier and target object identifier of the related edges that have been added, removed, or have changed content are extracted. The source object identifier and target object identifier together constitute the current defining endpoint pair of the cross-domain O&M related edges. The node corresponding to the source object identifier in the defining endpoint pair is denoted as the defining source endpoint, and the node corresponding to the target object identifier is denoted as the defining target endpoint. The node type of the defining source endpoint and the defining target endpoint is determined based on the value form of the source object identifier and the target object identifier: if the source object identifier or the target object identifier is a cross-domain O&M object identifier, then the corresponding defining endpoint is a cross-domain O&M object node, i.e., the corresponding node in the business acceptance object, rule processing object, channel interaction object, or result feedback object; if the source object identifier or the target object identifier is a combination of the data category identifier and the source domain identifier of the caliber coding parameter data record, then the corresponding defining endpoint is a caliber coding parameter data record node.

[0035] In the cross-domain operation and maintenance object association graph, taking the source endpoint and the target endpoint as the start and end nodes, the operation to extract the caliber coding parameter mapping and transmission path containing caliber coding parameter data is as follows: Taking the source endpoint as the start node and the target endpoint as the end node, in the directed graph structure of the cross-domain operation and maintenance object association graph, the path traversal operation is performed along the direction of the directed edge. The search scope of the path traversal operation is limited to the directed edges that involve caliber coding parameter data records at the source or target end of the directed edge, that is, the directed edges between the caliber coding parameter data records and the business acceptance objects, rule processing objects, channel interaction objects or result feedback objects, as well as the directed edges where there are field reference relationships between the caliber coding parameter data records themselves. The path traversal operation employs a depth-first traversal approach. Starting from the defined source endpoint, it expands progressively along the directed edges, recording all node sequences and directed edge sequences traversed along the path, until the defined target endpoint is reached or the traversal depth exceeds the maximum level depth of the directed edges in the cross-domain operation and maintenance object association graph for business processing order. For example, business acceptance objects, rule processing objects, channel interaction objects, and result feedback objects constitute four business processing stages, and the maximum level depth can be set to 4. All paths satisfying the search range constraints from the defined source endpoint to the defined target endpoint are used as the caliber coding parameter mapping and transmission path set. Each caliber coding parameter mapping and transmission path in the caliber coding parameter mapping and transmission path set contains a path node sequence, a path directed edge sequence, and a set of data category identifiers for the caliber coding parameter data carried in the path.

[0036] After determining the set of caliber coding parameter mapping and transmission paths, for each caliber coding parameter mapping and transmission path in the set, pre-constraint verification and post-constraint endpoint verification are performed. The object of the pre-constraint verification is the pre-concurrent cross-domain operation and maintenance object of the caliber coding parameter mapping and transmission path. The pre-concurrent cross-domain operation and maintenance object is defined as: in the path node sequence of the caliber coding parameter mapping and transmission path, all cross-domain operation and maintenance object nodes located before the defined source endpoint, that is, cross-domain operation and maintenance object nodes whose business processing sequence number is less than the business processing sequence number of the defined source endpoint in the path node sequence. If the defined source endpoint is a caliber coding parameter data record node rather than a cross-domain operation and maintenance object node, then the pre-concurrent cross-domain operation and maintenance object is all cross-domain operation and maintenance object nodes in the cross-domain operation and maintenance object association graph that have an association relationship with the defined source endpoint and have the smallest business processing sequence number. If there are no cross-domain operation and maintenance object nodes located before the defined source endpoint, then the set of pre-concurrent cross-domain operation and maintenance objects is empty, and the pre-constraint verification is determined to be a match.

[0037] The preceding constraint verification operation is as follows: For each preceding cross-domain operation and maintenance object in the preceding cross-domain operation and maintenance object set, a stable association edge is retrieved from the stable relationship benchmark set, where the cross-domain operation and maintenance object identifier of the preceding cross-domain operation and maintenance object is used as the source object identifier or the target object identifier. The retrieval operation is performed using the index established in step S2, which uses the combination of the source object identifier and the target object identifier as the retrieval key. If a stable association edge associated with the preceding cross-domain operation and maintenance object is found in the stable relationship benchmark set, and the data category identifier in the edge attribute set of the retrieved stable association edge is consistent with any data category identifier in the data category identifier set of the caliber coding parameter data carried in the caliber coding parameter mapping transmission path, then the current preceding cross-domain operation and maintenance object is determined to have passed the preceding constraint verification; if all preceding cross-domain operation and maintenance objects in the preceding cross-domain operation and maintenance object set pass the preceding constraint verification, then the preceding constraint verification of the current caliber coding parameter mapping transmission path is determined to be matched. If any of the preceding cross-domain operation and maintenance objects in the set of preceding cross-domain operation and maintenance objects does not have a stable association edge in the stable relationship benchmark set that matches the data category identifier of the caliber coding parameter data carried in the caliber coding parameter mapping and transmission path, then it is determined that the preceding constraint verification of the current caliber coding parameter mapping and transmission path does not match, the subsequent landing point verification is not performed on the current caliber coding parameter mapping and transmission path, and the current caliber coding parameter mapping and transmission path is skipped.

[0038] The execution object of the subsequent landing point verification is the subsequent cross-domain operation and maintenance object of the caliber coding parameter mapping and transmission path. The subsequent cross-domain operation and maintenance object is defined as: in the path node sequence of the caliber coding parameter mapping and transmission path, all cross-domain operation and maintenance object nodes located after the defined target endpoint, that is, cross-domain operation and maintenance object nodes whose business processing sequence number is greater than the business processing sequence number of the defined target endpoint in the path node sequence. If the defined target endpoint is a caliber coding parameter data record node rather than a cross-domain operation and maintenance object node, then the subsequent cross-domain operation and maintenance object is all the cross-domain operation and maintenance object nodes after the cross-domain operation and maintenance object node with the largest business processing sequence number that has an association relationship with the defined target endpoint in the cross-domain operation and maintenance object association graph. If there are no cross-domain operation and maintenance object nodes located after the defined target endpoint, then the set of subsequent cross-domain operation and maintenance objects is an empty set, the subsequent landing point verification is judged as a mismatch, and the path segment corresponding to the current caliber coding parameter mapping and transmission path is not included in the cross-domain operation and maintenance object affected segment set.

[0039] The subsequent landing point verification operation is as follows: For each subsequent cross-domain operation and maintenance object in the subsequent cross-domain operation and maintenance object set, retrieve the related edges in the relation change item set that have been added, deleted, or have changed in content, with the cross-domain operation and maintenance object identifier of the subsequent cross-domain operation and maintenance object as the source object identifier or target object identifier. The retrieval operation uses the cross-domain operation and maintenance object identifier of the subsequent cross-domain operation and maintenance object as the retrieval key to perform traversal matching in the relation change item set. If, in the set of relational change items, an associated edge that has been added, removed, or has undergone content changes is found that is related to subsequent cross-domain operation and maintenance objects, and the data category identifier in the edge attribute set of the retrieved associated edge that has been added, removed, or has undergone content changes matches any data category identifier in the data category identifier set of the caliber coding parameter data carried in the caliber coding parameter mapping and transmission path, then the current subsequent cross-domain operation and maintenance object is determined to have passed the subsequent landing point verification; if at least one subsequent cross-domain operation and maintenance object in the set of subsequent cross-domain operation and maintenance objects passes the subsequent landing point verification, then the subsequent landing point verification of the current caliber coding parameter mapping and transmission path is determined to be matched; if all subsequent cross-domain operation and maintenance objects in the set of subsequent cross-domain operation and maintenance objects fail the subsequent landing point verification, then the subsequent landing point verification of the current caliber coding parameter mapping and transmission path is determined to be mismatched, and the path segment corresponding to the current caliber coding parameter mapping and transmission path is not included in the cross-domain operation and maintenance object impact segment set.

[0040] The path segments corresponding to the caliber coding parameter mapping and transmission paths that match both the pre-constraint verification and the post-constraint endpoint verification are included in the cross-domain operation and maintenance object impact fragment set. A path segment is defined as: a sub-path extracted from the path node sequence and directed edge sequence of the caliber coding parameter mapping and transmission path, with the starting node being the defined source endpoint and the ending node being the defined target endpoint, bounded by the defined source endpoint and defined target endpoint. A path segment contains a complete set of records including the path segment node sequence, the path segment directed edge sequence, and the caliber coding parameter data carried within the path segment. Each record in the cross-domain operation and maintenance object impact fragment set contains the path segment node sequence, the path segment directed edge sequence, the complete set of records including the caliber coding parameter data carried within the path segment, the defined source endpoint identifier, the defined target endpoint identifier, the source object identifier and target object identifier of the stable associated edges corresponding to the pre-constraint verification matching, the source object identifier and target object identifier of the associated edges that have been added, removed, or have changed content corresponding to the post-constraint endpoint verification matching, and the execution timestamp of the path segment inclusion operation and the version identifier pair participating in the differential analysis. The cross-domain operation and maintenance object impact fragment set uses a combination of source endpoint identifier and target endpoint identifier as the retrieval key. A corresponding index is established in the version storage area of ​​the graph database to support efficient read operations when reading path segments from the cross-domain operation and maintenance object impact fragment set and performing semantic consistency verification with runtime representation data. After the cross-domain operation and maintenance object impact fragment set is written, the total number of records in the cross-domain operation and maintenance object impact fragment set and the version identifier pairs corresponding to the current preceding constraint verification and subsequent landing point verification operations are recorded and stored independently as cross-domain operation and maintenance object impact fragment set metadata.

[0041] Perform semantic consistency verification between the cross-domain operation and maintenance object impact fragment set and the operational characterization data, and extract the semantically abnormal fragment set, including: When reading path segments from the cross-domain operation and maintenance object impact fragment set, all records in the cross-domain operation and maintenance object impact fragment set are traversed. From each record, the path segment node sequence, the path segment directed edge sequence, and the complete set of records containing the caliber coding parameter data carried in the path segment are extracted. The path segment node sequence contains the starting node of the path segment (i.e., the defining source endpoint), the ending node of the path segment (i.e., the defining target endpoint), and all intermediate nodes traversed between the defining source endpoint and the defining target endpoint. The intermediate node types include cross-domain operation and maintenance object nodes and caliber coding parameter data record nodes. Each node in the path segment node sequence carries a node type identifier. When the node type identifier is a cross-domain operation and maintenance object node, the node carries the cross-domain operation and maintenance object identifier and the name of the business processing stage to which the cross-domain operation and maintenance object belongs; when the node type identifier is a caliber coding parameter data record node, the node carries a data category identifier and a source domain identifier.

[0042] The specific operations for matching link events, business result records, rule hit records, and operational alarm records in the operational representation data according to the cross-domain operation and maintenance objects and business processing order in the path segment are as follows: Extract all nodes whose node type identifier value is a cross-domain operation and maintenance object node from the path segment node sequence to form a path segment cross-domain operation and maintenance object node set; According to the business processing stage name of the cross-domain operation and maintenance object carried by each node in the path segment cross-domain operation and maintenance object node set, arrange the nodes in the path segment cross-domain operation and maintenance object node set according to the business processing order of business acceptance, rule processing, channel interaction, and result feedback to obtain an ordered sequence of path segment cross-domain operation and maintenance object nodes.

[0043] For cross-domain operation and maintenance object nodes in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment, whose business processing stage name is "business acceptance", the cross-domain operation and maintenance object identifier carried by the cross-domain operation and maintenance object node and the name of the business processing stage to which the cross-domain operation and maintenance object belongs are used as matching conditions. Link event records whose initiating business domain identifier is consistent with the business domain identifier corresponding to the cross-domain operation and maintenance object identifier are retrieved in the operation representation data. The retrieval operation uses the initiating business domain identifier as the retrieval key to perform traversal matching in the operation representation data. The retrieved link event records are used as the operation representation data that matches the cross-domain operation and maintenance object node corresponding to the business acceptance stage in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment. For cross-domain operation and maintenance object nodes in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment whose business processing stage name is rule processing, the cross-domain operation and maintenance object identifier carried by the cross-domain operation and maintenance object node and the business processing stage name to which the cross-domain operation and maintenance object belongs are used as matching conditions. The operation and representation data is searched for rule matching records in which the business domain identifier of the object matches the business domain identifier corresponding to the cross-domain operation and maintenance object identifier. The retrieved rule matching records are used as the operation and representation data that match the cross-domain operation and maintenance object nodes corresponding to the rule processing stage in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment. For cross-domain operation and maintenance object nodes in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment, whose business processing stage name is "channel interaction", the operation representation data is searched for link event records whose event type code belongs to cross-domain data transmission events and whose initiating end business domain identifier and receiving end business domain identifier are consistent with the business domain identifiers involved in the channel interaction stage. Additionally, business result records whose result type code belongs to tax-bank interaction receipt results and whose result-attributing business domain identifier is consistent with the business domain identifiers involved in the channel interaction stage are also searched. The retrieved link event records and business result records are used together as the operation representation data matching the cross-domain operation and maintenance object nodes corresponding to the channel interaction stage in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment. For cross-domain operation and maintenance object nodes in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment whose business processing stage name is "Result Feedback," the operation representation data is searched for business result records whose result type code belongs to tax declaration processing results, tax refund review results, or business anomaly return results, and whose result-attributed business domain identifier matches the business domain identifier involved in the result feedback stage. Operation alarm records whose alarm-attributed business domain identifier matches the business domain identifier involved in the result feedback stage are also searched. The retrieved business result records and operation alarm records are combined as operation representation data matching the cross-domain operation and maintenance object node corresponding to the result feedback stage in the ordered sequence of cross-domain operation and maintenance object nodes in the path segment. After completing the retrieval of operation representation data corresponding to each business processing stage, all matched link event records, business result records, rule hit records, and operation alarm records are summarized to form the set of operation representation data matching the current path segment.

[0044] When performing semantic consistency verification on the field meaning, business code attribution, and interface parameter value of the matched runtime representation data and the caliber coding parameter data in the path segment, the semantic consistency verification is performed sequentially in three dimensions: field meaning verification, business code attribution verification, and interface parameter value verification. The field meaning verification process is as follows: Extract records whose data category identifier is a tax caliber subset from the complete record set of caliber coding parameter data carried in the path segment. Extract the field meaning annotation field from each tax caliber subset record. The field meaning annotation field records the business semantic description of the caliber coding parameter field in structured text form. Extract rule-hitting records from the matched operational representation data set. Extract the caliber coding parameter data version number field used when hitting the rule in each rule-hitting record. Use the version number field as the version number index to retrieve the corresponding version number of the tax caliber subset record in the caliber coding parameter data. Extract the field meaning annotation field from the retrieved tax caliber subset record. Perform a field-by-field text comparison between the field meaning annotation fields of the tax caliber subset records in the path segment and the field meaning annotation fields of the tax caliber subset records associated with the rule-hitting records. If the text values ​​of the field meaning annotation fields of any corresponding fields are inconsistent, the field meaning verification is deemed inconsistent.

[0045] The operation for business code attribution verification is as follows: Extract records whose data category identifier value is a subset of the business code from the complete record set of caliber coding parameter data carried in the path segment. Take the local mapping relationship field and name conversion relationship field from each business code subset record. The local mapping relationship field records the mapping correspondence between the local code and the shared platform standard code, and the name conversion relationship field records the name conversion correspondence between the message field name and the shared platform standard code. Extract business result records from the matched operational representation data set. Take the result content field set from each business result record. The result content field set contains the actual business code values ​​used during business processing. Perform reverse mapping on the business code values ​​in the result content field set according to the local mapping relationship field and the name conversion relationship field in the business code subset record. Compare the shared platform standard code values ​​obtained after reverse mapping with the shared platform standard code values ​​in the business code subset records in the path segment one by one. If the shared platform standard code values ​​of any corresponding business code are inconsistent, it is determined that the business code attribution verification is inconsistent.

[0046] The interface parameter value verification process is as follows: Extract records from the complete record set of caliber coding parameter data carried in the path segment where the data category identifier is a subset of the interface parameters. Extract the parameter name, parameter data type, parameter value range, required parameter identifier, and default value definition fields from each interface parameter subset record. Extract link event records from the matched runtime representation data set. Extract the event type coding field and the receiving end business domain identifier field from each link event record. Using the receiving end business domain identifier field as the business domain range, search the caliber coding parameter data for interfaces where the interface call direction is the input parameter direction and the interface identifier matches the corresponding business domain identifier of the link event record. Interface parameter subset records: The actual call parameter values ​​corresponding to the link event record are compared with the parameter value range field of the interface parameter subset record in the path segment for range conformity verification. The verification method is to determine whether the actual call parameter value is within the value range defined by the parameter value range field of the interface parameter subset record, and at the same time, to check whether the interface parameter with the parameter mandatory identifier value exists in the actual call parameter set corresponding to the link event record. If the actual call parameter value of any interface parameter exceeds the value range defined by the parameter value range field, or if the interface parameter with the parameter mandatory identifier value does not have a corresponding value in the actual call parameter set, then the interface parameter value verification is determined to be inconsistent.

[0047] The semantic consistency review results of the current path segment are comprehensively judged based on three dimensions: field meaning review, business code attribution review, and interface parameter value review. If any one of the three is inconsistent, the current path segment is determined to be inconsistent and is included in the semantic anomaly fragment set. If all three are consistent, the current path segment is determined to be consistent and is not included in the semantic anomaly fragment set. Each record in the semantic anomaly fragment set contains the path segment node sequence of the inconsistent path segment, the directed edge sequence of the path segment, the complete set of caliber coding parameter data carried in the path segment, the source endpoint identifier, the target endpoint identifier, the set of inconsistent dimension identifiers, the inconsistent field name corresponding to each dimension in the inconsistent dimension identifier set, the caliber coding parameter data-side value and the runtime representation data-side value of the inconsistent field, as well as the execution timestamp of the semantic consistency review operation and the version identifier pair participating in the difference. The value range of the inconsistent dimension identifier set is any non-empty subset of the three inconsistencies: inconsistencies in field meaning review, inconsistencies in business code attribution review, and inconsistencies in interface parameter value review. The semantic anomaly fragment set uses the combination of the source endpoint identifier and the target endpoint identifier as the retrieval key, and establishes a corresponding index in the version storage area of ​​the graph database to support efficient reading operations when reading inconsistent path segments from the semantic anomaly fragment set and statistically analyzing the co-occurrence convergence characteristics of relational variation terms in the path segments. After the semantic anomaly fragment set is written, the total number of records in the semantic anomaly fragment set and the version identifier pair corresponding to this semantic consistency review operation are recorded and stored independently as semantic anomaly fragment set metadata.

[0048] Based on the co-occurrence convergence features of each relation variation term within the semantic anomaly fragment set, the relation drift source object is determined and the anomaly localization result is output, including: When reading inconsistent path segments from the semantic anomaly fragment set, the entire set of records in the semantic anomaly fragment set is traversed. From each record, the following are extracted: path segment node sequence, path segment directed edge sequence, complete record set of caliber coding parameter data carried in the path segment, source endpoint identifier, target endpoint identifier, and set of dimension identifiers indicating inconsistent verification. These fields together constitute a complete description of the current inconsistent path segment. Each node in the path segment node sequence carries a node type identifier. The node type identifier is determined by the following values: for cross-domain operation and maintenance object nodes, it carries the cross-domain operation and maintenance object identifier and the name of the business processing stage to which the cross-domain operation and maintenance object belongs; for caliber coding parameter data record nodes, it carries the data category identifier and source domain identifier. The path segment node sequence is arranged according to the business processing order of business acceptance, rule processing, channel interaction, and result feedback.

[0049] The operation of determining the set of relationship changes for path segment associations based on the relationship changes of path segment associations is as follows: Using the source endpoint identifier and target endpoint identifier of the path segment with inconsistent verification as search criteria, records whose source object identifier and target object identifier are consistent with the defined source endpoint identifier and the defined target endpoint identifier are retrieved from the drift candidate set of cross-domain operation and maintenance object association edges. From the retrieved drift candidate set records, the source object identifier, target object identifier, edge attribute set, and change type identifier of the association edges that have been added, deleted, or have changed content are extracted and included as the relationship changes for the current path segment associations with inconsistent verification. Simultaneously, using the cross-domain operation and maintenance object identifier carried by each cross-domain operation and maintenance object node in the path segment node sequence as the search key, records whose source object identifier or target object identifier is consistent with the cross-domain operation and maintenance object identifier of any cross-domain operation and maintenance object node in the path segment node sequence are searched in the relationship change item set. The associated edges that have been added, deleted, or have changed content in all the retrieved relationship change item set records are included in the relationship change item set of the path segment association that is inconsistent in the current review. Each relationship change item record in the relationship change item set carries a change type identifier, the source object identifier, the target object identifier, and the edge attribute set of the associated edge that has been added, deleted, or has changed content.

[0050] Based on the relationship changes associated with the path segment, cross-domain operation and maintenance objects, and business processing order, the operation of statistically analyzing the co-occurrence position, overlap range, and convergence order of relationship changes in the path segment is as follows: The co-occurrence position is defined as the landing position of each relationship change in the set of relationship changes in the path segment node sequence of the path segment where the current review is inconsistent. The landing position is represented by the arrangement number of the cross-domain operation and maintenance object node in the path segment node sequence that matches the source object identifier or target object identifier of the relationship change. The arrangement number starts from 1 and is incremented according to the business processing order of business acceptance, rule processing, channel interaction, and result feedback. If the node corresponding to the source object identifier and the node corresponding to the target object identifier of the relationship change both appear in the path segment node sequence, then the arrangement number of the node corresponding to the source object identifier is taken as the starting number of the co-occurrence position of the relationship change, and the arrangement number of the node corresponding to the target object identifier is taken as the ending number of the co-occurrence position of the relationship change. The overlapping range is defined as the sequence number interval of the path segment nodes where any two relational variables in the set of relational variables intersect or contain each other. The overlapping range is obtained by performing interval intersection calculation on the interval formed by the start and end numbers of the co-occurrence positions of all relational variables in the set of relational variables. The interval intersection calculation method is as follows: for each pair of relational variables in the set of relational variables, take the larger value of the start number of the co-occurrence positions of the two relational variables as the start number of the overlapping interval, and take the smaller value of the end number of the co-occurrence positions of the two relational variables as the end number of the overlapping interval. If the start number of the overlapping interval is less than or equal to the end number of the overlapping interval, it is determined that the two relational variables co-occur and overlap. The overlapping range is the sequence number interval of the path segment nodes from the start number of the overlapping interval to the end number of the overlapping interval. If multiple pairs of relational variables in the set of relational variables co-occur and overlap, then the union of all overlapping intervals is taken to obtain the complete overlapping range of the path segment where the current review is inconsistent. The convergence order is defined as the order in which each relation variable in the relation variable set is arranged in ascending order according to the starting index of its co-occurrence position. The convergence order is obtained by performing an ascending sort operation on the starting index of the co-occurrence position of all relation variables in the relation variable set. The comparison key of the sorting operation is the numerical value of the starting index of the co-occurrence position. Two relation variables with the same starting index of co-occurrence position are further arranged in a fixed priority order of content change, decrease, and increase according to the value of the change type identifier.The co-occurrence positions, overlapping ranges, and convergence orders of relational variables in the inconsistent path segments are combined to form the co-occurrence convergence features corresponding to the current inconsistent path segments. The co-occurrence convergence features are stored in the form of structured records. Each co-occurrence convergence feature record includes the source endpoint identifier and target endpoint identifier of the corresponding inconsistent path segment, the source object identifier and target object identifier combination of each relational variable in the relational variable set, the start and end sequence numbers of the co-occurrence positions corresponding to each relational variable, the start and end sequence numbers of the complete overlapping range, and the ordered sequence of relational variables arranged according to the convergence order. The co-occurrence convergence feature record also includes the execution timestamp of the co-occurrence convergence feature formation operation and the version identifier pair participating in the difference.

[0051] The operation to identify the cross-domain operation and maintenance object corresponding to the first relationship change item in the co-occurrence convergence feature that points to the inconsistency check as the relationship drift source object is as follows: For each co-occurrence convergence feature record, select the relationship change item with the convergence order number 1 from the ordered sequence of relationship change items arranged in convergence order, i.e., the relationship change item with the smallest co-occurrence position starting number, as the first relationship change item to enter the inconsistency check; if there are multiple relationship change items with the same co-occurrence position starting number and all of them are the minimum value, then select the relationship change item with the highest priority according to the priority order of content change, reduction, and increase change type identifiers. The first relationship change item to enter the review inconsistency is identified. From the source object identifier and target object identifier of the first relationship change item to enter the review inconsistency, the cross-domain operation and maintenance object identifier carried by the cross-domain operation and maintenance object node pointed to by the sequence number of the path segment node corresponding to the starting number of the co-occurrence position is taken. The cross-domain operation and maintenance object node is retrieved in the cross-domain operation and maintenance object association graph using the cross-domain operation and maintenance object identifier. The retrieved cross-domain operation and maintenance object node is identified as the relationship drift source object. The relationship drift source object carries the cross-domain operation and maintenance object identifier, the name of the business processing stage to which the cross-domain operation and maintenance object belongs, and the cross-domain operation and maintenance object creation timestamp.

[0052] After identifying the source object of the relationship drift, the output includes the source object, the relationship change item, and the path segment. The anomaly localization results are organized in structured record format. Each anomaly localization result record contains the following: the cross-domain operation and maintenance object identifier of the relationship drift source object, the name of the business processing stage to which the cross-domain operation and maintenance object belongs, and the creation timestamp of the cross-domain operation and maintenance object; the source object identifier, target object identifier, edge attribute set, and change type identifier of the first relationship change item directly associated with the relationship drift source object that entered the review inconsistency stage; and the path segment corresponding to the current co-occurring convergence feature that shows review inconsistency. The system includes: a sequence of path segment nodes, a sequence of directed edges in the path segment, a complete set of records of caliber coding parameter data carried in the path segment, identifiers defining the source endpoint, identifiers defining the target endpoint, and a set of dimension identifiers for verification inconsistencies; combinations of source and target object identifiers, start and end numbers of co-occurrence positions, start and end numbers of the complete overlapping range, and an ordered sequence of relationship changes arranged in convergence order in the relationship change set of the co-occurrence convergence feature record; and output timestamps and version identifier pairs participating in the differential analysis for the anomaly localization results. The anomaly localization results are persistently stored in a graph database, with an index built using the combination of the cross-domain operation and maintenance object identifier of the relationship drift source object and the version identifier pair participating in the differential analysis as the retrieval key, to support retrieval and aggregation statistics of anomaly localization results by relationship drift source object or differential batch. Repeatedly perform the co-occurrence convergence feature formation operation and the relation drift source object determination operation on all inconsistent path segments in the semantic anomaly fragment set until all records in the semantic anomaly fragment set have been processed. After all anomaly localization results are written to the graph database, record the total number of anomaly localization results and version identifier pairs corresponding to this operation, and store them independently as anomaly localization result metadata.

[0053] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0054] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0056] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0057] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0058] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0059] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, 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 medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0060] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0061] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A tax intelligent operation and maintenance management system based on a digital intelligence foundation, characterized in that, include: The graph construction module collects tax operation and maintenance related data from the shared platform, tax and banking channels, and internal multiple business domains, and constructs a cross-domain operation and maintenance object association graph according to the business processing order. Version Differentiation Module: Performs version solidification and adjacent version differential on the cross-domain operation and maintenance object association graph, and generates a set of relationship variation items and a set of stable relationship benchmarks for the corresponding caliber coding parameter data; Drift identification module: It compares and pairs the set of changing relation items with the set of stable relation baselines according to the field source, mapping level and transmission direction to identify the drift candidate set of cross-domain operation and maintenance object related edges; Impact verification module: Performs pre-constraint verification and post-landing point verification along the caliber encoding parameter mapping transmission path corresponding to the drift candidate set to form a cross-domain operation and maintenance object impact fragment set; Semantic verification module: Performs semantic consistency verification on the cross-domain operation and maintenance object impact fragment set and the operation representation data, and extracts the semantically abnormal fragment set; Source localization module: Based on the co-occurrence convergence features of each relation change term in the semantic anomaly fragment set, determine the relation drift source object and output the anomaly localization result.

2. The tax intelligent operation and maintenance management system based on a digital intelligence foundation according to claim 1, characterized in that, Collect tax operation and maintenance related data from the shared platform, tax and banking channels, and internal multiple business domains, and construct a cross-domain operation and maintenance object relationship graph according to the business processing order, specifically as follows: Tax calibers, business codes and interface parameters are read from the shared platform, tax-bank channels and internal multi-business domains and merged into caliber code parameter data. Link events, business result records, rule hit records and operation alarm records are read and merged into operation characterization data. Following the business processing sequence of business acceptance, rule processing, channel interaction, and result feedback, the caliber coding parameter data and operational characterization data are associated with the corresponding cross-domain operation and maintenance objects to generate a cross-domain operation and maintenance object association map.

3. The tax intelligent operation and maintenance management system based on a digital intelligence foundation according to claim 2, characterized in that, Version solidification and adjacent version differentiation are performed on the cross-domain operation and maintenance object relationship graph to generate a set of relationship variation items and a set of stable relationship benchmarks for the corresponding caliber coding parameter data. Specifically: The version identifier is the formation time and business processing order of the cross-domain operation and maintenance object association graph. The version is then fixed for the cross-domain operation and maintenance objects, caliber coding parameter data and associated edges in the cross-domain operation and maintenance object association graph. The associated edges of corresponding caliber coding parameter data in adjacent solidified versions are differentiated according to the source object, target object, and associated content; Add or remove related edges or change their content and categorize them into the relation change itemset. Consistent related edges are included in the stable relation benchmark set.

4. The tax intelligent operation and maintenance management system based on a digital intelligence foundation according to claim 3, characterized in that, The set of changing relation itemsets is paired with the set of stable relation baselines according to field source, mapping level, and propagation direction to identify the drift candidate set of cross-domain operation and maintenance object association edges, specifically: Extract the associated edges that have been added, removed, or have changed content from the relational change itemset; Extract stable relation edges from the stable relation benchmark set that have the same source object or the same target object as the associated edges; The associated edges are compared and paired with stable associated edges according to the field source, mapping level, and transmission direction. Associated edges with the same field source but different mapping levels or transmission directions are included in the drift candidate set of cross-domain operation and maintenance object associated edges.

5. A tax intelligent operation and maintenance management system based on a digital intelligence foundation as described in claim 4, characterized in that, Along the caliber encoding parameter mapping transmission path corresponding to the drift candidate set, perform pre-constraint verification and post-landing point verification to form a cross-domain operation and maintenance object impact fragment set, specifically: Read the cross-domain operation and maintenance object association edges from the drift candidate set, and extract the caliber coding parameter mapping transmission path containing caliber coding parameter data in the cross-domain operation and maintenance object association graph, using the source object and target object of the cross-domain operation and maintenance object association edge as the delimitation endpoints. Perform preceding constraint verification on the preceding cross-domain operation and maintenance object and the stable associated edge of the caliber coding parameter mapping and transmission path. Perform subsequent endpoint verification on the cross-domain operation and maintenance objects and relational change itemsets of the caliber coding parameter mapping and transmission path. Path segments that match both the preceding constraint check and the subsequent landing point check are categorized into the cross-domain operation and maintenance object impact segment set.

6. A tax intelligent operation and maintenance management system based on a digital intelligence foundation as described in claim 5, characterized in that, The semantic consistency of the cross-domain operation and maintenance object impact fragment set and the operational representation data is reviewed, and the semantically abnormal fragment set is extracted, specifically: Read path segments from the cross-domain operation and maintenance object impact fragment set, and match link events, business result records, rule hit records and operation alarm records in the operation representation data according to the cross-domain operation and maintenance objects and business processing order in the path segments; The matched operational representation data and the caliber coding parameter data in the path segment are semantically consistent in terms of field meaning, business coding affiliation, and interface parameter values. Path segments with inconsistent verification are classified into the semantic anomaly fragment set.

7. A tax intelligent operation and maintenance management system based on a digital intelligence foundation as described in claim 6, characterized in that, Based on the co-occurrence convergence features of each relation variation term within the semantic anomaly fragment set, the relation drift source object is determined and the anomaly localization result is output, specifically: Read inconsistent path segments from the set of semantically abnormal fragments, and according to the relationship changes associated with the path segments, cross-domain operation and maintenance objects and business processing order, count the co-occurrence position, overlap range and convergence order of the relationship changes in the path segments to form co-occurrence convergence features. The cross-domain operation and maintenance object corresponding to the first inconsistent relation change item pointed to in the co-occurrence convergence feature is identified as the relation drift source object, and the anomaly location result containing the relation drift source object, relation change item, and path segment is output.

Images