Data processing method and system
By integrating and quality-testing multi-source financial data, and building AI models and dual verification mechanisms, the problems of long business processing cycles and high labor costs in the financial system have been solved, achieving efficient and low-cost data processing.
Patent Information
- Application Number
- CN202511295287.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-11
- Publication Date
- 2025-12-23
AI Technical Summary
The existing financial system suffers from long business processing and decision-making cycles, high labor costs, and technical problems that make it difficult to control costs.
By accessing multi-source financial data, standardizing and integrating it, conducting quality testing and cleaning, building AI models for optimized training, establishing a dual verification mechanism, and constructing a distributed data repository, data traceability can be achieved.
It significantly shortens the business processing and decision-making cycle, effectively controls labor costs, and improves processing efficiency and accuracy.
Smart Images
Figure CN121190176A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of financial data processing, and in particular to a data processing method and system. BACKGROUND
[0002] With the deepening of the digital transformation of the financial industry, the business scale of financial subjects such as banks, payment institutions, and credit platforms continues to expand, and the amount of financial data generated daily grows exponentially, covering multi-dimensional data such as account information, transaction records, credit application materials, and customer behavior logs. Data types include structured numerical and textual data, as well as unstructured scanned copies and handwritten document images. Data sources involve bank core systems, third-party payment platforms, credit investigation agencies, and customer terminals, forming a complex financial data ecosystem.
[0003] Currently, in the current financial system data processing flow, a processing mode dominated by manual operation is generally relied on.
[0004] However, in financial system data processing, the business processing and decision-making cycle is long, and the labor cost ratio is high. In core financial businesses such as credit approval, transaction risk monitoring, and account reconciliation, manual data analysis, risk assessment, and result determination are required, which takes a long time. With the growth of business volume, labor costs rise linearly, making it difficult to control costs. SUMMARY
[0005] The purpose of the present application is to provide a data processing method and system to solve the technical problem of long business processing and decision-making cycle, high labor cost ratio, and difficulty in cost control in existing financial system data processing. In core financial businesses such as credit approval, transaction risk monitoring, and account reconciliation, manual data analysis, risk assessment, and result determination are required, which takes a long time. With the growth of business volume, labor costs rise linearly, making it difficult to control costs.
[0006] To achieve the above purpose, a data processing method is adopted, which includes the following steps:
[0007] Accessing multi-source financial data, collecting financial data, and standardizing and integrating financial data;
[0008] Quality detection and cleaning of data, completion of data quality optimization, and construction of a standardized data set;
[0009] Building an AI model, inputting the standardized data set into the AI model, and performing optimization training;
[0010] Inputting the fact financial business data to be processed into the model, analyzing the applicant information, performing transaction risk monitoring, and outputting the results;
[0011] Construct a dual verification mechanism to verify the output processing results;
[0012] Build a distributed data repository to receive verified processing results and establish financial business data traceability.
[0013] Among the steps involved are: accessing multi-source financial data, collecting financial data, and standardizing and integrating the financial data.
[0014] Financial data sources from the bank's core system, third-party payment platforms, credit reporting agency databases, and customer terminals are collected synchronously through the API gateway.
[0015] The system calls a pre-defined standardized processing algorithm to perform OCR recognition and format conversion on the collected unstructured data, and performs field mapping and data type unification on the structured data to eliminate format differences and semantic conflicts between different data sources. The unstructured data includes scanned copies of paper documents and images of handwritten signatures.
[0016] A correlation model is established based on the customer's unique identifier. Duplicate data from different data sources is deduplicated, and complementary data is concatenated to generate a complete customer profile.
[0017] Among the steps involved in data quality inspection and cleaning, data quality optimization, and the construction of a standardized dataset:
[0018] The isolated forest algorithm is used to scan the integrated dataset to identify outliers, missing values, and logically conflicting values, and to mark the location and type of outlier data.
[0019] For missing values, the KNN algorithm is used for interpolation and mode filling. For outliers, a secondary verification process is initiated, and the original data is automatically retrieved for correction. Logical conflict values are corrected according to preset business priorities.
[0020] The cleaned and standardized data were randomly divided into training, validation, and test sets in a 7:2:1 ratio, and stratified sampling was used to ensure that the business feature distribution of the three datasets was consistent.
[0021] In the step of building an AI model, the standardized dataset is input into the AI model for optimization training:
[0022] We chose the random forest algorithm to build the AI model;
[0023] The AI model is iteratively trained using the training set, and its hyperparameters, including the learning rate, number of iterations, and number of network layers, are adjusted in real time using the validation set. The trained model is then evaluated using the test set.
[0024] Among the steps, the process involves inputting the actual financial transaction data to be processed into the model, analyzing the applicant's information, monitoring transaction risks, and outputting the results:
[0025] By receiving financial business data in real time and automatically matching it with AI models corresponding to the business scenario;
[0026] The model is used to perform structured parsing of applicant information and time-series analysis of transaction data to identify abnormal transaction patterns.
[0027] Set up a risk matrix and combine it with the analysis results to automatically determine the business risk level;
[0028] The processing results are output in structured JSON format, including decision conclusions, key evidence, and risk scores.
[0029] In the step of constructing a dual verification mechanism to verify the output processing results:
[0030] The results output by the AI model are input into a preset rule engine, and the consistency of the results with financial regulatory requirements and internal business norms is compared to mark abnormal results that violate the rules.
[0031] 5% of the results that pass the rule engine verification are randomly sampled, and 100% of the marked abnormal results are extracted and assigned to the reviewers in the corresponding business areas. The AI analysis process and key data are displayed through a visual interface, and manual modification of conclusions is supported and the reasons for the correction are recorded.
[0032] Statistical analysis of AI model processing results includes independent processing pass rate, manual review and correction rate, and rule conflict rate, generating a verification report;
[0033] Mark the results that pass the verification and proceed to the next step. Update the status of the corrected results and store the revised examples in the model optimization sample library for subsequent model iterations.
[0034] Among the steps involved in building a distributed data repository, receiving verified processing results, and establishing financial business data traceability:
[0035] A multi-node distributed database is used to build the repository, and data partitions are divided according to business type;
[0036] Write the verified processing results into the repository according to the timestamp and business number indexing rules;
[0037] Build a traceability chain to generate blockchain transaction records from metadata. Each block contains the hash value of the previous block, making the data immutable and traceable.
[0038] It provides a visual traceability query interface, supports multi-dimensional retrieval by business number and customer identifier, displays the complete link from data collection to result output, and connects to the API interface of regulatory agencies to support the automatic export and reporting of compliance audit data.
[0039] This invention also provides a data processing system, including a data acquisition module, a data processing module, an AI model building module, an analysis module, a verification module, and a storage module; wherein:
[0040] The data acquisition module is used to access multi-source financial data, collect financial data, and standardize and integrate the financial data.
[0041] The data processing module is used to perform data quality detection and cleaning, optimize data quality, and construct a standardized dataset.
[0042] The AI model building module is used to build AI models by inputting standardized datasets into the AI models for optimization training.
[0043] The analysis module is used to input the actual financial business data to be processed into the model, analyze the applicant information, monitor transaction risks, and output results.
[0044] The verification module is used to construct a dual verification mechanism to verify the output processing results;
[0045] The storage module is used to build a distributed data repository, receive verified processing results, and establish financial business data traceability.
[0046] This invention discloses a data processing method and system that employs a data acquisition module, a data processing module, an AI model construction module, an analysis module, a verification module, and a storage module to perform the following steps: accessing multi-source financial data, collecting financial data, and standardizing and integrating the financial data; performing quality testing and cleaning on the data to optimize data quality and constructing a standardized dataset; constructing an AI model, inputting the standardized dataset into the AI model, and performing optimization training; inputting the actual financial business data to be processed into the model, analyzing applicant information, monitoring transaction risks, and outputting results; constructing a dual verification mechanism to verify the output processing results; and constructing a distributed data repository to receive the verified processing results and establish financial business data traceability. Through these methods, the time and manpower required for business processing and decision-making cycles are significantly shortened, and costs are effectively controlled. Attached Figure Description
[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0048] Figure 1 This is a flowchart of the data processing method of the present invention.
[0049] Figure 2 This is a flowchart of steps S100 of the present invention.
[0050] Figure 3 This is a flowchart of steps S200 of the present invention.
[0051] Figure 4 This is a flowchart of steps S300 of the present invention.
[0052] Figure 5 This is a flowchart of steps S400 of the present invention.
[0053] Figure 6 This is a flowchart of steps S500 of the present invention.
[0054] Figure 7 This is a flowchart of steps S600 of the present invention.
[0055] Figure 8 This is a schematic diagram of the data processing system of the present invention.
[0056] 701 - Data Acquisition Module, 702 - Data Processing Module, 703 - AI Model Building Module, 704 - Analysis Module, 705 - Verification Module, 706 - Storage Module. Detailed Implementation
[0057] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.
[0058] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.
[0059] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."
[0060] Please see Figures 1-7 The present invention provides a data processing method, comprising the following steps:
[0061] S100: Accesses multi-source financial data, collects financial data, and standardizes and integrates the financial data.
[0062] In this implementation, multi-source financial data is accessed, financial data is collected, and the financial data is standardized and integrated. The specific process is as follows:
[0063] S101: Synchronously collect financial data sources from the bank's core system, third-party payment platforms, credit reporting agency databases, and customer terminals through the API gateway;
[0064] S102: Call the preset standardized processing algorithm to perform OCR recognition and format conversion on the collected unstructured data, perform field mapping and data type unification on the structured data, and eliminate format differences and semantic conflicts between different data sources. Unstructured data includes scanned copies of paper documents and images of handwritten signatures.
[0065] S103: Establish an association model based on the customer's unique identifier, deduplicate data from different data sources, and concatenate complementary data fields to generate a complete customer profile.
[0066] In the above process, various financial data sources are collected synchronously from the bank's core system, third-party payment platforms, credit reporting agency databases, and customer terminals through the API gateway. Then, a preset standardized processing algorithm is invoked to perform OCR recognition and format conversion on the collected unstructured data (such as scanned copies of paper documents, handwritten signature images, etc.). Field mapping and data type unification operations are performed on structured data to eliminate format differences and semantic conflicts between different data sources. On this basis, an association model is built based on the customer's unique identifier to deduplicate data in different data sources. At the same time, complementary data is concatenated to generate a complete and accurate customer profile.
[0067] S200: Perform quality inspection and cleaning on the data, optimize the data quality, and build a standardized dataset.
[0068] In this implementation, the data undergoes quality inspection and cleaning to optimize data quality and a standardized dataset is constructed. The specific process is as follows:
[0069] S201: The isolated forest algorithm is used to scan the integrated dataset, identify outliers, missing values, and logically conflicting values, and mark the location and type of outlier data;
[0070] S202: Use the KNN algorithm to interpolate and fill missing values, and use the mode to fill missing values. Initiate a secondary verification process for outliers, automatically retrieve the original data from the source for correction, and correct logically conflicting values according to preset business priorities.
[0071] S203: The cleaned and standardized data are randomly divided into training set, validation set and test set in a ratio of 7:2:1, and stratified sampling is used to ensure that the business feature distribution of the three datasets is consistent.
[0072] In the above process, the Isolation Forest algorithm is used to conduct a comprehensive scan of the integrated dataset, accurately identifying outliers, missing values, and logically conflicting values, while clearly marking the location and specific type of outlier data. For missing values, the KNN algorithm is used for interpolation imputation (for continuous fields) and mode imputation (for discrete fields). For identified outliers, a secondary verification process is initiated, automatically retrieving the original data for correction. Logically conflicting values are corrected according to preset business priorities. After data cleaning and standardization, the data is randomly divided into training, validation, and test sets in a 7:2:1 ratio, and stratified sampling is used to ensure that the business feature distribution of the three datasets remains consistent.
[0073] S300: Build an AI model by inputting a standardized dataset into the AI model for optimized training.
[0074] In this embodiment, an AI model is constructed by inputting a standardized dataset into the AI model for optimization training. The specific process is as follows:
[0075] S301: Choose the random forest algorithm to build the AI model;
[0076] S302: Use the training set to iteratively train the AI model, adjust the model hyperparameters in real time using the validation set, including the learning rate, number of iterations, and number of network layers, and use the test set to evaluate the performance of the trained model.
[0077] In the above process, the random forest algorithm is selected to build the AI model. Then, the AI model is iteratively trained using the training set. The model hyperparameters, including the learning rate, number of iterations, and number of network layers, are adjusted in real time using the validation set. The performance of the trained model is evaluated using the test set.
[0078] S400: Input the actual financial transaction data to be processed into the model, analyze the applicant information, monitor transaction risks, and output the results.
[0079] In this implementation, the actual financial transaction data to be processed is input into the model, the applicant information is analyzed, transaction risk is monitored, and the results are output. The specific process is as follows:
[0080] S401: By receiving financial business data in real time, it automatically matches the corresponding business scenario with an AI model;
[0081] S402: Use models to perform structured parsing of applicant information and time-series analysis of transaction data to identify abnormal transaction patterns;
[0082] S403: Set up a risk matrix and automatically determine the business risk level based on the analysis results;
[0083] S404: Outputs the processing results in structured JSON format, including decision conclusions, key evidence, and risk scores.
[0084] In the above process, by receiving financial business data in real time, the system automatically and accurately matches it with the AI model of the corresponding business scenario; then, using the matched model, the applicant's information is structured and parsed, and time-series analysis is performed on the transaction data to identify abnormal transaction patterns; based on this, a risk matrix is set, and the risk level of the business is automatically determined in combination with the aforementioned analysis results. Specifically, the risk level of the business is automatically determined according to the preset risk matrix (levels 1 to 5). Level 1 (low risk) is directly approved, level 5 (high risk) is directly rejected, and risk warnings are generated for levels 2 to 4; finally, the system outputs the processing results in a structured JSON format, which includes the decision conclusion, key evidence, and key information on the risk score.
[0085] S500: Constructs a dual verification mechanism to verify the output processing results.
[0086] In this embodiment, a dual verification mechanism is constructed to verify the output processing result. The specific process is as follows:
[0087] S501: Input the results output by the AI model into the preset rule engine, compare the results with the consistency of financial regulatory requirements and internal business norms, and mark abnormal results that violate the rules;
[0088] S502: Randomly sample 5% of the results that pass the rule engine verification, extract 100% of the marked abnormal results, and assign them to the reviewers in the corresponding business areas. Display the AI analysis process and key data through a visual interface, support manual modification of conclusions and record the reasons for correction.
[0089] S503: Calculate the independent processing pass rate, manual review and correction rate, and rule conflict rate of the AI model processing results, and generate a verification report;
[0090] S504: Mark the results that pass the verification, proceed to the next step, update the status of the corrected results, and store the revised examples in the model optimization sample library for subsequent model iterations.
[0091] In the above process, the results output by the AI model are input into a preset rule engine and compared with financial regulatory requirements and internal business specifications for consistency, marking abnormal results that violate the rules. Results that pass the rule engine verification are randomly sampled at a rate of 5%, while marked abnormal results are sampled 100% and assigned to reviewers in the corresponding business areas. The AI analysis process and key data are displayed through a visual interface, allowing reviewers to manually modify conclusions and record the reasons for correction. Subsequently, indicators such as the independent processing pass rate, manual review and correction rate, and rule conflict rate of the AI model processing results are statistically analyzed, and a verification report is generated. Finally, the verified results are marked to proceed to the next step, the status of the corrected results is updated, and the modified examples are stored in the model optimization sample library for subsequent model iterations.
[0092] S600: Constructs a distributed data repository, receives verified processing results, and establishes financial business data traceability.
[0093] In this embodiment, a distributed data repository is constructed to receive verified processing results and establish financial business data traceability. The specific process is as follows:
[0094] S601: Uses a multi-node distributed database to build the repository and divides the data into partitions according to business type;
[0095] S602: Write the verified processing results into the repository according to the timestamp and business number indexing rules;
[0096] S603: Construct a traceability chain to generate blockchain transaction records from metadata. Each block contains the hash value of the previous block, making the data immutable and traceable.
[0097] S604: Provides a visual traceability query interface, supports multi-dimensional retrieval by business number and customer identifier, displays the complete link from data collection to result output, and connects to the API interface of regulatory agencies to support the automatic export and reporting of compliance audit data.
[0098] In the above process, a multi-node distributed database is used to build a repository. Data is partitioned according to business type, and the verified processing results are written into the repository according to the timestamp and business number index rules. At the same time, a traceability chain is built to generate blockchain transaction records from metadata, ensuring that each block contains the hash value of the previous block, thereby achieving the immutability and traceability of data. In addition, a visual traceability query interface is provided, which supports multi-dimensional retrieval based on business number, customer identifier and other dimensions, showing the complete link of data from collection to result output, and connecting to the API interface of regulatory agencies to realize the automatic export and reporting of compliance audit data.
[0099] Please see Figure 8 The present invention also provides a data processing system, including a data acquisition module 701, a data processing module 702, an AI model construction module 703, an analysis module 704, a verification module 705, and a storage module 706; wherein:
[0100] The data acquisition module 701 is used to access multi-source financial data, collect financial data, and standardize and integrate the financial data.
[0101] The data processing module 702 is used to perform data quality detection and cleaning, optimize data quality, and construct a standardized dataset.
[0102] The AI model building module 703 is used to build an AI model by inputting a standardized dataset into the AI model for optimization training.
[0103] The analysis module 704 is used to input the financial transaction data to be processed into the model, analyze the applicant information, monitor transaction risks, and output results.
[0104] The verification module 705 is used to construct a dual verification mechanism to verify the output processing results;
[0105] The storage module 706 is used to build a distributed data repository, receive the verified processing results, and establish financial business data traceability.
[0106] In this embodiment, the data acquisition module 701 accesses multi-source financial data, collects financial data, and standardizes and integrates the financial data. The data processing module 702 performs quality detection and cleaning on the data, optimizes the data quality, and constructs a standardized dataset. The AI model construction module 703 constructs an AI model, inputs the standardized dataset into the AI model, and performs optimization training. The analysis module 704 inputs the factual financial business data to be processed into the model, analyzes the applicant information, monitors transaction risks, outputs results, and constructs a dual verification mechanism. The verification module 705 verifies the output processing results. The storage module 706 constructs a distributed architecture data repository, receives the verified processing results, and establishes financial business data traceability. Through the above methods, the time and manpower required for business processing and decision-making cycles can be greatly shortened, and costs can be effectively controlled.
[0107] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.
[0108] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.
Claims
1. A data processing method, characterized in that, Includes the following steps: Access multi-source financial data, collect financial data, and standardize and integrate the financial data; Perform quality inspection and cleaning on the data, optimize the data quality, and build a standardized dataset; Build an AI model by inputting a standardized dataset into the AI model for optimized training; Input the actual financial transaction data to be processed into the model, analyze the applicant information, monitor transaction risks, and output the results. Construct a dual verification mechanism to verify the output processing results; Build a distributed data repository to receive verified processing results and establish financial business data traceability.
2. The data processing method as described in claim 1, characterized in that, In the steps of accessing multi-source financial data, collecting financial data, and standardizing and integrating financial data: Financial data sources from the bank's core system, third-party payment platforms, credit reporting agency databases, and customer terminals are collected synchronously through the API gateway. The system calls a pre-defined standardized processing algorithm to perform OCR recognition and format conversion on the collected unstructured data, and performs field mapping and data type unification on the structured data to eliminate format differences and semantic conflicts between different data sources. The unstructured data includes scanned copies of paper documents and images of handwritten signatures. A correlation model is established based on the customer's unique identifier. Duplicate data from different data sources is deduplicated, and complementary data is concatenated to generate a complete customer profile.
3. The data processing method as described in claim 1, characterized in that, In the steps of data quality inspection and cleaning, data quality optimization, and construction of standardized datasets: The isolated forest algorithm is used to scan the integrated dataset to identify outliers, missing values, and logically conflicting values, and to mark the location and type of outlier data. For missing values, the KNN algorithm is used for interpolation and mode filling. For outliers, a secondary verification process is initiated, and the original data is automatically retrieved for correction. Logical conflict values are corrected according to preset business priorities. The cleaned and standardized data were randomly divided into training, validation, and test sets in a 7:2:1 ratio, and stratified sampling was used to ensure that the business feature distribution of the three datasets was consistent.
4. The data processing method as described in claim 1, characterized in that, In the process of building an AI model, inputting a standardized dataset into the AI model and performing optimization training: We chose the random forest algorithm to build the AI model; The AI model is iteratively trained using the training set, and its hyperparameters, including the learning rate, number of iterations, and number of network layers, are adjusted in real time using the validation set. The trained model is then evaluated using the test set.
5. The data processing method as described in claim 1, characterized in that, In the steps of inputting the actual financial transaction data to be processed into the model, analyzing applicant information, monitoring transaction risks, and outputting results: By receiving financial business data in real time and automatically matching it with AI models corresponding to the business scenario; The model is used to perform structured parsing of applicant information and time-series analysis of transaction data to identify abnormal transaction patterns. Set up a risk matrix and combine it with the analysis results to automatically determine the business risk level; The processing results are output in structured JSON format, including decision conclusions, key evidence, and risk scores.
6. The data processing method as described in claim 1, characterized in that, In the step of constructing a dual verification mechanism to verify the output processing results: The results output by the AI model are input into a preset rule engine, and the consistency of the results with financial regulatory requirements and internal business norms is compared to mark abnormal results that violate the rules. 5% of the results that pass the rule engine verification are randomly sampled, and 100% of the marked abnormal results are extracted and assigned to the reviewers in the corresponding business areas. The AI analysis process and key data are displayed through a visual interface, and manual modification of conclusions is supported and the reasons for the correction are recorded. Statistical analysis of AI model processing results includes independent processing pass rate, manual review and correction rate, and rule conflict rate, generating a verification report; Mark the results that pass the verification and proceed to the next step. Update the status of the corrected results and store the revised examples in the model optimization sample library for subsequent model iterations.
7. The data processing method as described in claim 1, characterized in that, In the steps of building a distributed data repository, receiving verified processing results, and establishing financial business data traceability: A multi-node distributed database is used to build the repository, and data partitions are divided according to business type; Write the verified processing results into the repository according to the timestamp and business number indexing rules; Build a traceability chain to generate blockchain transaction records from metadata. Each block contains the hash value of the previous block, making the data immutable and traceable. It provides a visual traceability query interface, supports multi-dimensional retrieval by business number and customer identifier, displays the complete link from data collection to result output, and connects to the API interface of regulatory agencies to support the automatic export and reporting of compliance audit data.
8. A data processing system, applied to the data processing method as described in claim 1, characterized in that, It includes a data acquisition module, a data processing module, an AI model building module, an analysis module, a verification module, and a storage module; among which: The data acquisition module is used to access multi-source financial data, collect financial data, and standardize and integrate the financial data. The data processing module is used to perform data quality detection and cleaning, optimize data quality, and construct a standardized dataset. The AI model building module is used to build AI models by inputting standardized datasets into the AI models for optimization training. The analysis module is used to input the actual financial business data to be processed into the model, analyze the applicant information, monitor transaction risks, and output results. The verification module is used to construct a dual verification mechanism to verify the output processing results; The storage module is used to build a distributed data repository, receive verified processing results, and establish financial business data traceability.