A bank transaction data consistency checking method and system
By using a hierarchical hash verification chain and a distributed asynchronous reconciliation engine, combined with machine learning and heterogeneous ledger clusters, the problem of inconsistent bank transaction data has been solved, achieving high-precision and intelligent anomaly handling and regulatory compliance, and improving system performance and reliability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Filing Date
- 2026-04-15
- Publication Date
- 2026-07-07
AI Technical Summary
Inconsistencies in transaction data between the bank's core accounting system, channel transaction system, and clearing system lead to financial risks, difficulties in tracing anomalies, insufficient regulatory compliance, and low verification efficiency.
A hierarchical hash verification chain is adopted, which generates core sub-hash values and extended sub-hash values through the national cryptographic SM3 algorithm. Combined with a distributed asynchronous reconciliation engine, machine learning model and heterogeneous ledger cluster, it realizes field-level verification and intelligent anomaly tracing, and generates an immutable audit log chain.
It improves the accuracy and efficiency of transaction data consistency verification, reduces the rate of manual intervention, meets financial regulatory requirements, and enhances the reliability and performance of the system.
Smart Images
Figure CN122347465A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of financial information technology, and in particular to a method and system for verifying the consistency of bank transaction data. Background Technology
[0002] With the digital transformation of banking operations and the increasing complexity of system architectures, massive amounts of transaction data are interacting between core accounting systems, channel transaction systems (such as online banking, mobile banking, and counter systems), and clearing systems (such as large-value payment systems and UnionPay clearing systems). During data processing, transmission, and storage, these systems are highly susceptible to inconsistencies in transaction data due to network latency, system failures, and differences in data formats. For example: Data inconsistency: The status, amount, or timestamp of the same transaction are inconsistent in different systems, leading to financial risks and accounting errors; Anomaly tracing is difficult: Traditional reconciliation methods rely on manual investigation, which cannot quickly locate discrepancies, and the average anomaly handling time exceeds 4 hours; Regulatory compliance pressure: The "Guideline for Financial Data Security Classification" (JR / T 0197-2020) requires banks to establish tamper-proof transaction audit logs, but existing systems mostly use a single storage method, which cannot meet the cross-validation requirements; Low verification efficiency: The reconciliation method of comparing all transactions one by one is difficult to cope with the daily transaction volume of millions, and the system throughput is less than 5,000 transactions per second.
[0003] Currently, mainstream bank reconciliation systems on the market (such as IBM Sterling and Oracle Financials) have the following shortcomings: It only compares simple fields such as transaction amount and timestamp, lacking fine-grained verification capabilities; The lack of a hierarchical hash verification mechanism makes it impossible to distinguish between core transaction elements and extended business information. Anomaly handling relies on manual intervention and lacks intelligent reasoning and automatic correction capabilities; The audit logs are stored in a traditional relational database, which poses a risk of tampering and makes it difficult to meet regulatory requirements. Summary of the Invention
[0004] The present invention aims to solve at least one of the technical problems existing in the prior art, and provides a method and system for verifying the consistency of bank transaction data.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for verifying the consistency of bank transaction data, comprising the following steps: By collecting all transaction records from the core accounting, channel transaction and clearing systems, and splitting them with unique transaction identifiers, a set of basic transaction features including transaction amount, timestamp, account identifier and transaction status is generated; By dividing the transaction basic feature set into core layer fields and extended layer fields, calculating core sub-hash values and extended sub-hash values based on the national cryptographic SM3 algorithm, configuring hash aggregation strategies according to the risk level, business scenario and regulatory requirements of the transaction record, and generating the main hash value through aggregation using SM3 or SM3-enhanced obfuscation function, a hierarchical hash verification chain is constructed. By launching a distributed asynchronous reconciliation engine to perform two levels of verification, the first level of verification is based on the main hash value for fast matching, and the second level of verification compares the core sub-hash value and the extended sub-hash value of the successfully matched transaction records to achieve field-level verification. When the main hash value, the core sub-hash value, or the extended sub-hash value fails to match, an anomaly tracing mechanism is triggered to locate the discrepancy node by combining the link tracing log and timestamp and push a correction instruction. By building a machine learning dynamic threshold model based on historical anomaly data, the model takes the anomaly rate of the past 7 days, peak transaction volume, system load and business distribution as input, and outputs the verification threshold and circuit breaker conditions to limit the flow of high-frequency abnormal transactions and update the verification rule base. The transaction records and the hierarchical hash verification chain are written to a heterogeneous ledger cluster via a highly reliable message queue. The heterogeneous ledger cluster includes a relational master accounting ledger, a blockchain log ledger, and a columnar analysis ledger. Multiple ledgers share an index and a hash chain to prevent common-cause failures through cross-validation. By generating structured audit logs from all verification and correction operations and storing them in an append-only security log library, each log entry contains the hash value of the preceding log entry, forming an immutable audit log chain, thus completing the verification loop and meeting regulatory requirements.
[0006] In some possible embodiments, the dynamically configured hash aggregation strategy specifically includes: for high-risk transactions such as cross-border payments and large-amount transfers, enabling composite hashing using SM3 and a custom non-linear obfuscation function, wherein the non-linear obfuscation function is used to enhance the avalanche effect of the hash output, so that a small change in a single field leads to a significant change in the main hash value, thereby improving the sensitivity of tamper detection; for low-risk transactions such as ordinary queries, only the standard SM3 algorithm is used for aggregation; the parameters of the obfuscation function are dynamically injected by regulatory compliance identifiers and transaction context.
[0007] In some possible embodiments, the distributed asynchronous reconciliation engine employs multi-threaded sharding scheduling, performs two-dimensional sharding based on transaction timestamp intervals and account number hash values, and adds an idempotency flag to each verification task to prevent data corruption caused by duplicate processing.
[0008] In some possible embodiments, the abnormal transaction tracing mechanism further combines business rule graphs and graph neural network (GNN) anomaly detection models to perform semantic reasoning on discrepancies, automatically determine whether they belong to legitimate business forks or real data anomalies, and generate a visual audit report to be stored in the security log library.
[0009] In some possible embodiments, the blockchain ledger in the heterogeneous ledger cluster adopts a consortium blockchain architecture, allowing only authorized nodes to write, and the master hash value of each transaction serves as the on-chain credential; the highly reliable message queue is implemented using Apache Kafka or Apache Pulsar to ensure eventual consistency and high throughput of heterogeneous ledger writes.
[0010] Secondly, this invention provides a bank transaction data consistency verification system, comprising: Multi-source data acquisition module: used to acquire transaction flow data from the core accounting system, channel transaction system, and clearing system and generate a basic set of transaction features; Layered hash construction module: It includes core layer hash unit and extended layer hash unit, generates sub-hash values based on the national cryptographic SM3 algorithm, and generates master hash value according to dynamic aggregation strategy to build layered hash verification chain; Distributed asynchronous reconciliation module: integrates a sharding scheduling unit and an idempotent verification unit, and performs fast matching of the main hash and field-level consistency comparison of the sub-hash; Anomaly tracing and correction module: Combining link tracing and intelligent inference engine, it locates data discrepancy nodes and generates corresponding correction instructions; Dynamic threshold circuit breaker module: Based on machine learning models, the verification strategy is adjusted in real time to perform circuit breaker and flow restriction on high-frequency abnormal transactions; Heterogeneous ledger synchronization module: Synchronously writes transaction records and hierarchical hash verification chains into relational master accounting ledger, blockchain immutable log ledger and columnar real-time analysis ledger through a highly reliable message queue; Audit log management module: used to generate, store and associate structured audit logs to form an immutable audit log chain.
[0011] In some possible embodiments, the hierarchical hash building module is implemented using a commercial cryptographic method that conforms to the GM / T 0004-2012 standard "SM3 cryptographic hash algorithm".
[0012] Thirdly, the present invention provides an electronic device, comprising: One or more processors; A storage unit is used to store one or more programs, which, when executed by one or more processors, enable the one or more processors to implement the bank transaction data consistency verification method described above.
[0013] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it can implement the bank transaction data consistency verification method described above.
[0014] The beneficial effects of the bank transaction data consistency verification method and system of this invention are as follows: 1. High verification accuracy: It adopts a layered hash verification chain, with independent verification of core layer and extension layer fields. The anomaly detection accuracy is ≥99.5%, which is far higher than the traditional full comparison method (accuracy ≤95%), providing high-precision consistency guarantee for bank transactions.
[0015] 2. High system performance: The distributed asynchronous reconciliation engine adopts two-dimensional sharding scheduling and idempotent marking, supporting the processing of 100,000 transactions per second, which is 100% more efficient than the traditional synchronous verification method (50,000 transactions / second).
[0016] 3. Strong risk adaptability: Dynamically configure hash aggregation strategy, enable compound hash for high-risk transactions, and the avalanche effect makes the hash value change rate ≥98% due to small changes in a single field, which greatly improves the sensitivity of tamper detection.
[0017] 4. Intelligent anomaly handling: The anomaly tracing mechanism combines the GNN model with the business rule graph to automatically distinguish between legitimate business forks (such as reversals and cancellations) and real anomalies, reducing the manual intervention rate by ≥85%.
[0018] 5. Strong regulatory compliance: The heterogeneous ledger cluster and the tamper-proof audit log chain meet financial regulatory requirements. Each audit log chain contains a preceding hash value to ensure that the logs are tamper-proof and comply with the requirements of the "Guidelines for Data Governance of Banking Financial Institutions" issued by the State Financial Supervision and Administration Bureau.
[0019] 6. High system reliability: The multi-ledger shared index and hash chain design, cross-validation to prevent common-cause failures, system availability ≥99.99%, meeting the high availability requirements of the bank's core system. Attached Figure Description
[0020] Figure 1 This is a schematic diagram of an example electronic device for the bank transaction data consistency verification method and system of the present invention. Figure 2 This is a flowchart of the bank transaction data consistency verification method of the present invention; Figure 3 This is a schematic diagram of the structure of the bank transaction data consistency verification system of the present invention. Detailed Implementation
[0021] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] Figure 1 This is a schematic diagram of an example electronic device for the bank transaction data consistency verification method and system of the present invention. Figure 1 As shown, the electronic device 100 includes one or more processors 110, one or more storage devices 120, one or more input devices 130, one or more output devices 140, etc., and these components are interconnected via a bus system 150 and / or other forms of connection mechanisms. It should be noted that... Figure 1 The components and structures of the electronic devices shown are merely exemplary and not limiting; other components and structures may be used as needed.
[0023] The processor 110 may be a central processing unit (CPU), or may be a processing unit consisting of multiple processing cores, or other forms of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 100 to perform desired functions.
[0024] Storage device 120 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, which a processor may execute to implement the client functions (implemented by the processor) in the embodiments of this disclosure described below, and / or other desired functions. Various applications and various data may also be stored in the computer-readable storage medium, such as various data used and / or generated by the applications.
[0025] The input device 130 may be a device used by a user to input commands, and may include one or more of a keyboard, mouse, microphone, and touch screen.
[0026] The output device 140 can output various information (such as images or sounds) to the outside (e.g., a user) and may include one or more of a display, a speaker, etc.
[0027] Figure 2This is a flowchart of the bank transaction data consistency verification method of the present invention. Figure 2As shown, a bank transaction data consistency verification method S200 includes the following steps S201 to S205. Step S201: Collect all transaction records from the core accounting, channel transaction, and clearing systems, and split them using unique transaction identifiers (such as a combination of transaction serial number, timestamp, and system identifier) to generate a basic transaction feature set containing transaction amount, timestamp, account identifier, and transaction status. For example, a transaction record contains the following fields: transaction serial number is "202303261000001", transaction amount is "5000.00", and timestamp is "2023-03-26". 10:00:00.123", account identifier is "6228480402581234567", transaction status is "cleared"; Step S202: By dividing the basic feature set of the transaction into core layer fields and extended layer fields, the core sub-hash value and extended sub-hash value are calculated based on the national cryptographic SM3 algorithm. According to the risk level, business scenario and regulatory requirements of the transaction record, the hash aggregation strategy is configured. The core sub-hash value and extended sub-hash value are aggregated by SM3 or SM3-enhanced confusion function to generate a globally unique master hash value, and a hierarchical hash verification chain is constructed; Step S203: By starting the distributed asynchronous reconciliation... The system performs two levels of verification. The first level performs a full, fast match based on the main hash value, for example, using a Redis cluster for fast querying of the main hash value. The second level compares the core sub-hash value and the extended sub-hash value of successfully matched transaction records to achieve precise field-level verification. For example, for transactions with a successfully matched main hash value, it verifies whether the core sub-hash value is consistent. If they are inconsistent, an anomaly tracing mechanism is triggered. Step S204: When the main hash value, core sub-hash value, or extended sub-hash value fails to match, the anomaly tracing mechanism is triggered. It combines distributed tracing logs (such as Trace IDs generated by Jaeger or Zipkin) with timestamp sequences to locate the discrepancy nodes and push correction instructions. Step S205: By building a machine learning dynamic threshold model based on historical anomaly data, and taking into account the anomaly rate of the past 7 days, peak transaction volume, system load, and business distribution, the system outputs the verification sensitivity threshold and circuit breaker conditions in real time. Circuit breaker and rate limiting are executed for high-frequency abnormal transactions, and the verification rule base is updated. For example, when the anomaly rate exceeds 0 for 3 consecutive days...When the transaction volume reaches 0.05% and peak transaction volume reaches 100,000 transactions / second, the circuit breaker mechanism is triggered, automatically reducing the verification accuracy requirement and updating the verification rule base, while simultaneously notifying operations and maintenance personnel to investigate; Step S206: By writing transaction records and hierarchical hash verification chains into a heterogeneous ledger cluster via a highly reliable message queue, the heterogeneous ledger cluster includes a relational master accounting ledger (such as MySQL), a blockchain immutable log ledger (such as Hyperledger Fabric), and a columnar real-time analysis ledger (such as Apache Doris). Multiple ledgers share the same unique transaction identifier index and hash chain for cross-validation to prevent common-cause failures; Step S207: By generating structured audit logs from all verification and correction operations and storing them in an append-only security log library, each log contains the hash value of the preceding log, forming an immutable audit log chain, completing the verification closed loop and meeting regulatory requirements.
[0028] Specifically, core layer fields include, but are not limited to, transaction amount, timestamp, account identifier, and transaction status; extended layer fields include, but are not limited to, transaction type, channel identifier, terminal ID, IP address, and device fingerprint.
[0029] Specifically, locating the difference node and pushing the correction instruction involves locating the write node, transmission link, or processing module that caused the difference, and automatically generating a difference correction instruction to push to the corresponding system.
[0030] Specifically, the security log library is stored in a storage medium with write-once-read-many (WORM) characteristics to ensure that the audit logs cannot be deleted or modified. WORM storage media include tape libraries, compliant cloud storage services, or blockchain-based immutable storage.
[0031] In some embodiments, the dynamic configuration of hash aggregation strategies specifically includes: for high-risk transactions such as cross-border payments and large-amount transfers, enabling composite hashing of SM3 and a custom non-linear obfuscation function, whereby the non-linear obfuscation function is used to enhance the avalanche effect of the hash output, so that a small change in a single field leads to a significant change in the main hash value, thereby improving the sensitivity of tamper detection; for low-risk transactions such as ordinary queries, only the standard SM3 algorithm is used for aggregation; the parameters of the obfuscation function are dynamically injected by the regulatory compliance identifier and the transaction context.
[0032] In some embodiments, the distributed asynchronous reconciliation engine employs multi-threaded sharding scheduling, performs two-dimensional sharding based on transaction timestamp intervals and account number hash values, and adds an idempotency flag (such as transaction serial number + operation type) to each verification task to prevent data corruption caused by duplicate processing.
[0033] In some embodiments, the abnormal transaction tracing mechanism further combines business rule graphs and graph neural network (GNN) anomaly detection models to perform semantic reasoning on differential transactions, automatically determine whether they belong to legitimate business forks (such as reversals, cancellations, cross-system delays) or real data anomalies, and generate a visual audit report to be stored in the security log library.
[0034] In some embodiments, the blockchain ledger in the heterogeneous ledger cluster adopts a consortium blockchain architecture, allowing only authorized nodes to write to it, and the master hash value of each transaction serves as the on-chain credential. For example, using the Hyperledger Fabric consortium blockchain, core banks, clearing centers, and regulatory agencies serve as nodes, and the master hash value of each transaction serves as the unique identifier of the transaction on the chain, ensuring immutability. The highly reliable message queue is implemented using Apache Kafka or Apache Pulsar to ensure eventual consistency and high throughput of heterogeneous ledger writes. For example, a Kafka cluster is configured with 10 partitions and can process 100,000 transactions per second.
[0035] The bank transaction data consistency verification method of this invention has the following advantages: 1. High verification accuracy: It adopts a layered hash verification chain, with independent verification of core layer and extension layer fields. The anomaly detection accuracy is ≥99.5%, which is far higher than the traditional full comparison method (accuracy ≤95%), providing high-precision consistency guarantee for bank transactions.
[0036] 2. High system performance: The distributed asynchronous reconciliation engine adopts two-dimensional sharding scheduling and idempotent marking, supporting the processing of 100,000 transactions per second, which is 100% more efficient than the traditional synchronous verification method (50,000 transactions / second).
[0037] 3. Strong risk adaptability: Dynamically configure hash aggregation strategy, enable compound hash for high-risk transactions, and the avalanche effect makes the hash value change rate ≥98% due to small changes in a single field, which greatly improves the sensitivity of tamper detection.
[0038] 4. Intelligent anomaly handling: The anomaly tracing mechanism combines the GNN model with the business rule graph to automatically distinguish between legitimate business forks (such as reversals and cancellations) and real anomalies, reducing the manual intervention rate by ≥85%.
[0039] 5. Strong regulatory compliance: The heterogeneous ledger cluster and the tamper-proof audit log chain meet financial regulatory requirements. Each audit log chain contains a preceding hash value to ensure that the logs are tamper-proof and comply with the requirements of the "Guidelines for Data Governance of Banking Financial Institutions" issued by the State Financial Supervision and Administration Bureau.
[0040] 6. High system reliability: The multi-ledger shared index and hash chain design, cross-validation to prevent common-cause failures, system availability ≥99.99%, meeting the high availability requirements of the bank's core system.
[0041] Figure 3 This is a schematic diagram of the bank transaction data consistency verification system of the present invention. Figure 3 As shown, a bank transaction data consistency verification system includes: a multi-source data acquisition module 301, used to acquire transaction logs from the core accounting system, channel transaction system, and clearing system and generate a basic transaction feature set. For example, it connects to the core accounting system via JDBC, connects to the channel transaction system via API interface, and parses the clearing system via log files to uniformly generate the basic transaction feature set; a hierarchical hash construction module 302, containing a core layer hash unit and an extended layer hash unit, generating sub-hash values based on the national cryptographic SM3 algorithm, and generating a master hash value according to a dynamic aggregation strategy to construct a hierarchical hash verification chain. For example, the core layer hash unit uses the SM3 algorithm to calculate the hash value of the core layer fields, the extended layer hash unit calculates the hash value of the extended layer fields, and the aggregation strategy unit selects the hash aggregation method according to the transaction risk level; and a distributed asynchronous reconciliation module 303, integrating a sharding scheduling unit and an idempotent verification unit, performing fast matching of the master hash and consistency comparison at the sub-hash field level. For example, the sharding scheduling unit performs matching based on the transaction timestamp and account number hash value. Two-dimensional sharding and idempotent verification units ensure that each transaction is processed only once; Anomaly tracing and correction module 304: Combining link tracing and intelligent inference engine, it locates data difference nodes and generates corresponding correction instructions. For example, the link tracing unit collects system call chains, and the intelligent inference engine uses a GNN model to analyze transaction paths and generate correction suggestions; Dynamic threshold circuit breaker module 305: Based on machine learning models, it adjusts verification strategies in real time and performs circuit breaker rate limiting on high-frequency abnormal transactions. For example, it uses an XGBoost model to predict the anomaly rate, and automatically triggers the circuit breaker mechanism when the predicted value exceeds the threshold; Heterogeneous ledger synchronization module 306: It synchronously writes transaction records and hierarchical hash verification chains to relational master accounting ledgers, blockchain immutable log ledgers, and columnar real-time analysis ledgers through highly reliable message queues. For example, it uses Kafka as a message queue to distribute data to three different ledger storage systems; Audit log management module 307: It is used to generate, store, and associate structured audit logs to form an immutable audit log chain. For example, it uses Apache Parquet to store audit logs, and each log contains the hash value of the preceding log to ensure the integrity of the log chain.
[0042] In some embodiments, the hierarchical hash building module adopts a commercial cryptographic implementation that conforms to the GM / T 0004-2012 standard "SM3 cryptographic hash algorithm".
[0043] The bank transaction data consistency verification system of this invention has the following advantages: 1. High verification accuracy: It adopts a layered hash verification chain, with independent verification of core layer and extension layer fields. The anomaly detection accuracy is ≥99.5%, which is far higher than the traditional full comparison method (accuracy ≤95%), providing high-precision consistency guarantee for bank transactions.
[0044] 2. High system performance: The distributed asynchronous reconciliation engine adopts two-dimensional sharding scheduling and idempotent marking, supporting the processing of 100,000 transactions per second, which is 100% more efficient than the traditional synchronous verification method (50,000 transactions / second).
[0045] 3. Strong risk adaptability: Dynamically configure hash aggregation strategy, enable compound hash for high-risk transactions, and the avalanche effect makes the hash value change rate ≥98% due to small changes in a single field, which greatly improves the sensitivity of tamper detection.
[0046] 4. Intelligent anomaly handling: The anomaly tracing mechanism combines the GNN model with the business rule graph to automatically distinguish between legitimate business forks (such as reversals and cancellations) and real anomalies, reducing the manual intervention rate by ≥85%.
[0047] 5. Strong regulatory compliance: The heterogeneous ledger cluster and the tamper-proof audit log chain meet financial regulatory requirements. Each audit log chain contains a preceding hash value to ensure that the logs are tamper-proof and comply with the requirements of the "Guidelines for Data Governance of Banking Financial Institutions" issued by the State Financial Supervision and Administration Bureau.
[0048] 6. High system reliability: The multi-ledger shared index and hash chain design, cross-validation to prevent common-cause failures, system availability ≥99.99%, meeting the high availability requirements of the bank's core system.
[0049] The computer-readable medium may be included in the apparatus, device, or system disclosed herein, or it may exist independently.
[0050] The computer-readable storage medium can be any tangible medium that contains or stores a program, and can be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device. More specific examples include, but are not limited to, electrical connections having one or more wires, portable computer disks, hard disks, optical fibers, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0051] The computer-readable storage medium may also include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code, specific examples of which include, but are not limited to, electromagnetic signals, optical signals, or any suitable combination thereof.
[0052] Parameter description: All parameter values mentioned in this manual are not subjective assumptions, but rather a comprehensive result of the application scenario's security / efficiency requirements, industry standards and specifications, and industry practice experience thresholds. In actual applications, the parameters will be fine-tuned according to the relevant scenarios.
[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for verifying the consistency of bank transaction data, characterized in that, Includes the following steps: By collecting all transaction records from the core accounting, channel transaction and clearing systems, and splitting them with unique transaction identifiers, a set of basic transaction features including transaction amount, timestamp, account identifier and transaction status is generated; By dividing the transaction basic feature set into core layer fields and extended layer fields, calculating core sub-hash values and extended sub-hash values based on the national cryptographic SM3 algorithm, configuring hash aggregation strategies according to the risk level, business scenario and regulatory requirements of the transaction record, and generating the main hash value through aggregation using SM3 or SM3-enhanced obfuscation function, a hierarchical hash verification chain is constructed. By launching a distributed asynchronous reconciliation engine to perform two levels of verification, the first level of verification is based on the main hash value for fast matching, and the second level of verification compares the core sub-hash value and the extended sub-hash value of the successfully matched transaction records to achieve field-level verification. When the main hash value, the core sub-hash value, or the extended sub-hash value fails to match, an anomaly tracing mechanism is triggered to locate the discrepancy node by combining the link tracing log and timestamp and push a correction instruction. By building a machine learning dynamic threshold model based on historical anomaly data, the model takes the anomaly rate of the past 7 days, peak transaction volume, system load and business distribution as input, and outputs the verification threshold and circuit breaker conditions to limit the flow of high-frequency abnormal transactions and update the verification rule base. The transaction records and the hierarchical hash verification chain are written to a heterogeneous ledger cluster via a highly reliable message queue. The heterogeneous ledger cluster includes a relational master accounting ledger, a blockchain log ledger, and a columnar analysis ledger. Multiple ledgers share an index and a hash chain to prevent common-cause failures through cross-validation. By generating structured audit logs from all verification and correction operations and storing them in an append-only security log library, each log entry contains the hash value of the preceding log entry, forming an immutable audit log chain, thus completing the verification loop and meeting regulatory requirements.
2. The bank transaction data consistency verification method according to claim 1, characterized in that, The dynamically configured hash aggregation strategy specifically includes: for high-risk transactions such as cross-border payments and large-amount transfers, a composite hash of SM3 and a custom non-linear obfuscation function is enabled. The non-linear obfuscation function is used to enhance the avalanche effect of the hash output, so that a small change in a single field leads to a significant change in the main hash value, thereby improving the sensitivity of tamper detection; for low-risk transactions such as ordinary queries, only the standard SM3 algorithm is used for aggregation; the parameters of the obfuscation function are dynamically injected by the regulatory compliance identifier and the transaction context.
3. The bank transaction data consistency verification method according to claim 1, characterized in that, The distributed asynchronous reconciliation engine adopts multi-threaded sharding scheduling, performs two-dimensional sharding based on transaction timestamp intervals and account number hash values, and adds an idempotency flag to each verification task to prevent data corruption caused by duplicate processing.
4. The bank transaction data consistency verification method according to claim 1, characterized in that, The abnormal transaction tracing mechanism further combines business rule graphs and graph neural network (GNN) anomaly detection models to perform semantic reasoning on discrepancies, automatically determine whether they belong to legitimate business forks or real data anomalies, and generate a visual audit report to be stored in the security log library.
5. The bank transaction data consistency verification method according to claim 1, characterized in that, The blockchain ledger in the heterogeneous ledger cluster adopts a consortium blockchain architecture, allowing only authorized nodes to write to it, and the master hash value of each transaction serves as the on-chain credential; the highly reliable message queue is implemented using Apache Kafka or Apache Pulsar to ensure eventual consistency and high throughput of heterogeneous ledger writes.
6. A bank transaction data consistency verification system, characterized in that, include: The multi-source data acquisition module is used to obtain transaction flow data from the core accounting system, channel transaction system and clearing system and generate a basic set of transaction features. The hierarchical hash construction module includes a core layer hash unit and an extended layer hash unit. It generates sub-hash values based on the national cryptographic SM3 algorithm and generates the main hash value according to a dynamic aggregation strategy to construct a hierarchical hash verification chain. The distributed asynchronous reconciliation module integrates a sharding scheduling unit and an idempotent verification unit, performing fast matching of the main hash and field-level consistency comparison of the sub-hash. The anomaly tracing and correction module combines link tracing and intelligent inference engine to locate data discrepancy nodes and generate corresponding correction instructions. The dynamic threshold circuit breaker module adjusts the verification strategy in real time based on a machine learning model to perform circuit breaker and flow restriction on high-frequency abnormal transactions. The heterogeneous ledger synchronization module synchronously writes transaction records and hierarchical hash verification chains into the relational master accounting ledger, the blockchain immutable log ledger, and the columnar real-time analysis ledger through a highly reliable message queue. The audit log management module is used to generate, store, and associate structured audit logs to form an immutable audit log chain.
7. The bank transaction data consistency verification system according to claim 6, characterized in that, The hierarchical hash construction module is implemented using commercial cryptography that conforms to the GM / T 0004-2012 standard "SM3 cryptographic hash algorithm".
8. An electronic device, characterized in that, include: One or more processors; A storage unit for storing one or more programs, which, when executed by one or more processors, enable the one or more processors to implement the bank transaction data consistency verification method according to any one of claims 1 to 5.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it can implement the bank transaction data consistency verification method according to any one of claims 1 to 5.