Data transaction-oriented dynamic compliance checking method and evaluation system

Abstract

The invention discloses a data transaction-oriented dynamic compliance check method and evaluation system, and the method comprises the steps: generating an initial credit rating and a preliminary compliance strategy in combination with the type of transaction data and the historical transaction behaviors and enterprise qualifications of both parties, and carrying out the online learning and incremental learning of the strategy according to the latest compliance rule standard and laws and regulations. Carrying out online training and obtaining dynamic compliance evaluation information; after each transaction is completed, the model collects orders, delivery information and settlement data in real time, generates data transaction vouchers, and feeds back the transaction vouchers to the model through continuously generated transaction vouchers, so that credit rating and compliance check strategies of both parties are dynamically adjusted, and it is ensured that the transaction meets compliance requirements. According to the dynamic compliance evaluation model, the admission threshold and the compliance cost are effectively reduced by flexibly adjusting the compliance strategy, and the dynamic compliance evaluation model adapts to the rapid change of the data transaction market.

Description

Technical Field

[0001] The present invention belongs to the technical field of data inspection, and particularly relates to a dynamic compliance inspection method and evaluation system for data transactions. Background Art

[0002] With the rapid development of the data trading market, compliance inspection has become a key link to ensure the legality, transparency, and security of data transactions. Most of the existing compliance inspection methods are based on static rules and are difficult to adapt to the rapid changes in regulations. Especially in complex scenarios such as cross-border transactions and financial data transactions, the traditional compliance inspection methods are inefficient and prone to repeated verification, resulting in high entry barriers and increased compliance costs. In addition, the existing models cannot dynamically update credit ratings, resulting in the inability to identify potential risks of both trading parties in real time, increasing compliance risks and costs. To solve the above problems, the present invention provides a dynamic compliance inspection method and model for data transactions, which can flexibly adjust compliance strategies, lower entry barriers, and improve compliance efficiency. Summary of the Invention

[0003] To solve the deficiencies of the prior art and achieve the purpose of reducing the entry barriers and compliance costs of data transactions and adapting to the rapid changes in the data trading market, the present invention adopts the following technical solutions: A dynamic compliance inspection method for data transactions includes the following steps: Step 1: Real-time collect the transaction data of both trading parties, and classify the transaction data based on the corresponding compliance rules according to both trading parties and their respective industry fields; the transaction data includes order information, delivery status, and clearing and settlement, and the classification of the data includes data type, data source, and sensitivity; Step 2: Analyze the transaction data and extract the key information related to the transaction process; Step 3: Input the classified transaction data and the key information related to the transaction process parsed out into a dynamic compliance model, and generate a preliminary compliance strategy according to the preset compliance rules and laws and regulations; the dynamic compliance model is obtained by online training through combining the latest laws and regulations standards and using online learning and incremental learning strategies, obtains the classified transaction data, and applies corresponding compliance strategies to different categories of transaction data; Extract key information such as transaction amount, transaction frequency, and default records from the transaction process in the historical transaction records; obtain the qualification information of the transaction entity, and integrate the key information and qualification information into the data for the qualification assessment basis of the transaction entity; use data cleaning tools to clean the historical transaction data, including removing outliers, handling missing data, and standardizing the data to ensure accurate analysis and training of the machine learning model; for unstructured data sources (such as contracts or reports), use natural language processing techniques to extract key information such as the key financial status and risk records for the qualification assessment of the entity, which is used for the generation of compliance strategies. Step 4: Based on the generated preliminary compliance strategy, dynamically analyze the historical transaction data and current transaction behaviors, evaluate the effectiveness of the compliance strategy and adjust the compliance strategy. Combine the historical transaction records and the qualification information of the transaction entity to generate an initial credit rating; after each transaction is completed, real-time collect the order, delivery information, and clearing and settlement data during the transaction process to generate a unique transaction voucher, which contains key information such as the order number, transaction amount, and clearing and settlement status; dynamically adjust the credit ratings of both parties to the transaction according to the transaction voucher; analyze each transaction behavior and its compliance result to obtain a compliance trend prediction for future transactions; based on the compliance result and prediction result, conduct an adjustment analysis of the current compliance strategy to evaluate whether the current strategy adapts to future trends; according to the analysis result, formulate a compliance strategy update plan and adjust the compliance strategies of both parties to the transaction in future transactions accordingly.

[0004] Further, in step 4, after each transaction is completed, a transaction voucher is generated. The specific process of generating the transaction voucher includes: real-time obtaining the order information, delivery information, and clearing and settlement data during the transaction process from the transaction model through an API interface or a message queue (such as Kafka or RabbitMQ); preprocessing the collected transaction process data, including format standardization, duplicate removal, and data cleaning, to ensure the integrity and consistency of the data; integrating the transaction process data into structured data containing key information of the transaction process, including order number, transaction amount, delivery status, identity information of both parties to the transaction, and key information of clearing and settlement information; generating a unique transaction voucher based on the structured data. The transaction voucher includes core information such as transaction time, amount, and credit rating. The transaction voucher will be used as the input basis for subsequent dynamic adjustment of the credit rating and optimization of the compliance strategy to ensure the integrity and traceability of each transaction behavior.

[0005] Furthermore, in step 4, the generated transaction vouchers are used to dynamically adjust the credit ratings and compliance strategies of both trading parties, which specifically includes: using the transaction vouchers generated after each transaction as input, where the vouchers contain data such as order numbers, transaction amounts, clearing and settlement status, delivery information, and the historical credit ratings of both trading parties; using machine learning algorithms (such as random forests or support vector machines) to analyze the data in each transaction voucher, dynamically updating the credit ratings of both trading parties, and the update of the credit rating is based on factors such as performance, order success rate, and timeliness of clearing and settlement; based on the updated credit ratings, re-evaluating the risk levels of both trading parties, and the risk levels directly affect the compliance strategies for subsequent transactions; for high-risk transactions, implementing a more stringent compliance review process, while for low-risk transactions, simplifying the compliance process; through incremental learning, continuously accumulating transaction voucher data and optimizing the compliance strategy to ensure that each transaction process complies with the latest compliance requirements and credit rating standards.

[0006] Furthermore, in step 1, regular expressions are used to parse structured or semi-structured transaction data; for unstructured transaction data (such as documents, reports, etc.), natural language processing (NLP) is performed. Specifically, a word vector model is used to preprocess the text to extract the context information and semantic content of the dataset; based on named entity recognition (NER), important information about both trading parties and their respective domains in the transaction data is identified, such as data types (such as financial data, medical data, government data, social media data, etc.), the identity information of the data provider and recipient, etc., for data classification; through dependency syntax analysis, the syntactic structure in the data description is parsed to extract the relationships between important information for data classification to ensure the accuracy of classification.

[0007] Furthermore, in step 2, natural language processing is performed on the transaction data to parse the transaction data that generates order information, delivery data, and clearing and settlement data during the transaction process, and then parse it to extract key information, such as transaction amount, identity information of both trading parties, etc.

[0008] Furthermore, in step 2, a multi-level data extraction method is adopted to parse and process the transaction data, which specifically includes: First, regular expressions are used to parse structured and semi-structured data to extract key fields in the transaction process, such as order numbers, transaction amounts, and dates; for unstructured text data, a word vector model (such as Word2Vec or GloVe) is used to convert the text into vector form to capture semantic information; Then, based on deep learning named entity recognition (NER), key information related to the transaction process is identified, including the identity information of both trading parties, transaction amount, and date, etc. Dependency syntactic analysis is used to parse complex sentence structures to ensure the logical relationships of the identified key information in the context; Finally, the extracted key information is verified through a vector similarity algorithm (such as cosine similarity) to ensure that the data is consistent with the expected format and type.

[0009] A dynamic compliance evaluation system for data transactions, comprising a data classification module, a data extraction module, a compliance policy generation module, a policy evaluation and adjustment module, characterized in that: The data classification module collects the transaction data of both trading parties, and classifies the transaction data based on the corresponding compliance rules according to the trading parties and their respective industry fields; The data extraction module parses the transaction data, extracts the key information related to the transaction process, including structured and unstructured data such as order numbers, transaction amounts, identity information, etc., and provides it to the subsequent compliance policy generation module to ensure the integrity and consistency of the transaction data; The compliance policy generation module inputs the classified transaction data and the key information related to the transaction process parsed out into a dynamic compliance model, and generates a preliminary compliance policy according to the preset compliance rules and laws and regulations; The policy evaluation and adjustment module dynamically analyzes the historical transaction data and the current transaction behavior based on the generated preliminary compliance policy, evaluates the effectiveness of the compliance policy and adjusts the compliance policy.

[0010] Furthermore, the policy evaluation and adjustment module includes a transaction voucher generation module and a credit rating and compliance adjustment module; The transaction voucher generation module, after each transaction is completed, collects in real time the order, delivery information and clearing and settlement data of the transaction process, generates a unique transaction voucher, the voucher contains key information such as order number, transaction amount, clearing and settlement status, etc., integrates these data into structured information, and the voucher serves as the basic record of transaction compliance to ensure traceability; The credit rating and compliance adjustment module combines the historical transaction records and the qualification information of the trading entities to generate an initial credit rating; dynamically adjusts the credit ratings of both trading parties according to the transaction vouchers; analyzes each transaction behavior and its compliance result to obtain the compliance trend prediction of future transactions; based on the compliance result and the prediction result, conducts an adjustment analysis on the current compliance policy to evaluate whether the current policy adapts to future trends; according to the analysis result, formulates a compliance policy update plan and adjusts the compliance policies of both trading parties in future transactions accordingly.

[0011] The advantages and beneficial effects of the present invention are as follows: Based on the latest regulatory standards, the present invention continuously optimizes and online trains through machine learning and incremental learning strategies, can adapt to regulatory changes in real time, and ensures the accuracy and timeliness of compliance inspections for data transactions. At the same time, the model can dynamically adjust credit ratings and compliance strategies, reduce manual intervention, and improve the efficiency and flexibility of compliance audits in the data transaction process. By automatically generating transaction vouchers and dynamically adjusting compliance strategies, the model can effectively reduce compliance costs and enhance the security and reliability of transactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a flowchart of the method in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] The following further details the specific embodiments of the present invention with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for the purpose of illustrating and explaining the present invention and are not intended to limit the present invention.

[0014] As Figure 1 shown, a dynamic compliance inspection method for data transactions includes the following steps: Step 1: Real-time collect the transaction data of both parties to the transaction, and classify the transaction data based on the corresponding compliance rules according to the parties to the transaction and their respective industry fields. In an embodiment of the present invention, by real-time collecting data such as order information, delivery status, and clearing and settlement of both parties to the transaction, and classifying the data based on dimensions such as data type, source, and sensitivity, the specific classification process is as follows: Use regular expressions to parse structured or semi-structured transaction data, such as keyword fields such as dataset name, data source, data format, and data size; for unstructured transaction data (such as documents, reports, etc.), through natural language processing (NLP) technology, specifically use word vector models to preprocess the text and extract the context information and semantic content of the dataset; based on named entity recognition (NER) technology, important information in the transaction data can be identified, such as data type (such as financial data, medical data, government data, social media data, etc.), identity information of the data provider and recipient, etc. Dependency syntactic analysis is further used to parse the syntactic structure in the data description and extract the relationships between key information to ensure the accuracy of classification.

[0015] The classification basis includes but is not limited to dimensions such as data type (such as financial data, medical data, social data, etc.), data source (such as enterprises, governments, individuals), and data sensitivity (such as public data, sensitive data, confidential data).

[0016] Step 2: Use natural language processing technology to parse transaction data, collect transaction data such as order information, delivery data, and clearing and settlement data generated during the transaction process, and parse it to extract key information related to the transaction process, such as transaction amount, identity information of both parties to the transaction, etc.

[0017] Adopt a multi-level data extraction technology to parse and process the data generated during the transaction process, specifically including: First, parse structured and semi-structured data through regular expressions to extract key fields such as order numbers, transaction amounts, and dates; for unstructured text data, use word vector models (such as Word2Vec or GloVe) to convert the text into vector form to capture semantic information; Then, based on the named entity recognition (NER) technology of deep learning, identify important contents such as the identity information of both parties to the transaction, transaction amount, and date; dependency syntactic analysis is used to parse complex sentence structures to ensure the logical relationship of the identified key information in the context; Finally, verify the extracted information through a vector similarity algorithm (such as cosine similarity) to ensure that the data is consistent with the expected format and type.

[0018] Step 3: Input the classified transaction data and the key information related to the transaction process parsed into a dynamic compliance model, and generate a preliminary compliance strategy according to the preset compliance rules and laws and regulations.

[0019] The dynamic compliance model is obtained through online training by combining the latest laws and regulations standards and using online learning and incremental learning strategies. It obtains the classified transaction data and applies corresponding compliance strategies to different types of transaction data. For example, financial data and medical data need to comply with specific industry regulations, while public social data may require lower compliance standards.

[0020] The dynamic compliance model generates a preliminary compliance strategy through historical transaction data, specifically including: extracting key information such as transaction amount, transaction frequency, and default records from historical transaction records in the data warehouse through SQL query language or NoSQL database; through API interfaces, the model automatically obtains enterprise qualification information, such as government registration information, credit reports, etc., and integrates these data as the basis for enterprise qualification assessment; using data cleaning tools to clean historical transaction data, including removing outliers, handling missing data, and standardizing the data to ensure accurate analysis and training of machine learning models; for unstructured data sources (such as contracts or reports), use natural language processing technology to extract key financial conditions, risk records, etc. key information to support the generation of compliance strategies.

[0021] Step 4: Based on the generated preliminary compliance strategy, dynamically analyze historical transaction data and current transaction behaviors, evaluate the effectiveness of the compliance strategy, and adjust the compliance strategy. Specifically, the model generates an initial credit rating by combining historical transaction records and enterprise qualifications; after each transaction is completed, it real-time collects order, delivery information, and clearing and settlement data of the transaction to generate a unique transaction voucher, which contains key information such as order number, transaction amount, and clearing and settlement status; the model dynamically adjusts the credit ratings of both parties according to the transaction voucher; the model analyzes each transaction behavior and its compliance result to obtain a prediction of the compliance trend of future transactions; based on the compliance result and prediction result, it conducts an adjustment analysis on the current compliance strategy to evaluate whether the current strategy adapts to future trends; according to the analysis result, it formulates a compliance strategy update plan and adjusts the compliance strategies of both parties in future transactions accordingly.

[0022] The model generates a transaction voucher after each transaction is completed. The specific process of generating the transaction voucher includes: obtaining order information, delivery information, and clearing and settlement data from the transaction model in real time through an API interface or a message queue (such as Kafka or RabbitMQ); preprocessing the collected transaction data, including format standardization, deduplication, and data cleaning, to ensure the integrity and consistency of the data; integrating the preprocessed data into structured data, which contains key information such as order number, transaction amount, delivery status, identity information of both parties to the transaction, and clearing and settlement information; generating a unique transaction voucher based on this data, and the voucher contains core information such as transaction time, amount, and credit rating. The transaction voucher will be used as the input basis for subsequent dynamic adjustment of credit ratings and optimization of compliance strategies to ensure the integrity and traceability of each transaction behavior.

[0023] The model uses the generated transaction voucher to dynamically adjust the credit ratings and compliance strategies of both parties to the transaction. Specifically, it includes: taking the transaction voucher generated after each transaction is completed as the input, and the voucher contains data such as order number, transaction amount, clearing and settlement status, delivery information, and historical credit ratings of both parties to the transaction; using machine learning algorithms (such as random forest or support vector machine) to analyze the data in each transaction voucher and dynamically update the credit ratings of both parties to the transaction. The update of the credit rating is based on factors such as performance, order success rate, and timeliness of clearing and settlement; according to the updated credit rating, the model re-evaluates the risk levels of both parties to the transaction, and the risk level directly affects the compliance strategy of subsequent transactions; for high-risk transactions, the model will implement a more stringent compliance review process, while for low-risk transactions, it simplifies the compliance process; through incremental learning, the model continuously accumulates transaction voucher data and optimizes the compliance strategy to ensure that each transaction process meets the latest compliance requirements and credit rating standards.

[0024] A dynamic compliance evaluation system for data transactions, comprising: a data classification module, a data extraction module, a compliance policy generation module, a transaction voucher generation module, and a credit rating and compliance adjustment module.

[0025] The data classification module classifies the parsed data. According to dimensions such as data type, source, and sensitivity, the model applies corresponding compliance policies to different categories of data to ensure that the data complies with relevant laws and regulations. The data extraction module is used to extract key information from the transaction process. The model extracts structured and unstructured data such as order numbers, transaction amounts, and identity information and provides it to the subsequent compliance policy generation module to ensure the integrity and consistency of transaction data. The compliance policy generation module generates compliance policies based on the extracted data and historical records. The model combines historical transaction data, enterprise qualification information, and current transaction data and generates compliance inspection policies through machine learning analysis. The transaction voucher generation module generates a unique transaction voucher after each transaction is completed. The model real-time collects order, delivery, and clearing and settlement data, integrates this data into structured information, and the voucher serves as the basic record of transaction compliance to ensure traceability. The credit rating and compliance adjustment module dynamically adjusts the credit rating and compliance policies using the generated transaction vouchers. The model updates the credit rating by analyzing the performance and order success rate and adjusts the compliance policies according to the rating, and conducts more stringent audits on high-risk transactions.

[0026] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements 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 dynamic compliance checking method for data transactions, characterized in that The steps include: Step 1: Collect transaction data of both parties in real time and classify the transaction data according to the parties and their industry sectors based on the corresponding compliance rules; Step 2: Analyze the transaction data and extract key information related to the transaction process; Step 3: Input the classified transaction data and the key information related to the analyzed transaction process into the dynamic compliance model, and generate a preliminary compliance strategy based on the preset compliance rules; The dynamic compliance model generates preliminary compliance strategies through historical transaction data, including: extracting key information of the transaction process in the historical transaction records from the database; obtaining the qualification information of the transaction entity, integrating the key information and qualification information into data for the qualification evaluation of the transaction entity; extracting the entity qualification evaluation information for the generation of compliance strategies; Step 4: Based on the generated preliminary compliance strategy, dynamically analyze historical transaction data and current transaction behavior, evaluate the effectiveness of the compliance strategy and adjust the compliance strategy; Generate an initial credit rating by combining historical transaction records and transaction entity qualification information; collect transaction process data in real time after each transaction is completed to generate a unique transaction voucher; dynamically adjust the credit rating of both parties to the transaction based on the transaction voucher; analyze each transaction behavior and its compliance results to obtain compliance trend forecasts for future transactions; adjust and analyze the current compliance strategy based on the compliance results and forecast results to evaluate whether the current strategy is suitable for future trends; formulate a compliance strategy update plan based on the analysis results, and adjust the compliance strategies of both parties to the transaction in future transactions accordingly.

2. A dynamic compliance checking method for data transactions according to claim 1, characterized in that: In step 4, a transaction voucher is generated after each transaction is completed. The transaction voucher generation process specifically includes: acquiring transaction process data in real time; integrating the transaction process data into structured data containing key information of the transaction process, including order number, transaction amount, delivery status, identity information of both parties to the transaction, and key information of clearing and settlement information; generating a unique transaction voucher based on the structured data, and the transaction voucher includes core information such as transaction time, amount, and credit rating.

3. A dynamic compliance checking method for data transactions according to claim 1, characterized in that: In step 4, the generated transaction voucher is used to dynamically adjust the credit rating and compliance strategy of both parties to the transaction, specifically including: taking the transaction voucher generated after each transaction is completed as input, the voucher contains data such as order number, transaction amount, clearing and settlement status, delivery information, and historical credit ratings of both parties to the transaction; using a machine learning algorithm to analyze the data in each transaction voucher, and dynamically updating the credit ratings of both parties to the transaction, and the update of the credit rating is based on performance, order success rate, and timeliness of clearing and settlement; re-evaluating the risk level of both parties to the transaction based on the updated credit rating, and the risk level directly affects the compliance strategy of subsequent transactions; implementing a stricter compliance review process for high-risk transactions, and simplifying the compliance process for low-risk transactions; and continuously accumulating transaction voucher data and optimizing compliance strategies through incremental learning.

4. A dynamic compliance checking method for data transactions according to claim 1, characterized in that: In step 1, regular expressions are used to parse structured or semi-structured transaction data; natural language processing is performed on unstructured transaction data to extract context information and semantic content of the data set; Based on named entity recognition, important information about the two parties to the transaction and their fields in the transaction data is identified for data classification. Through dependency syntax analysis, the syntactic structure in the data description is parsed and the relationship between important information is extracted for data classification.

5. A dynamic compliance checking method for data transactions according to claim 1, characterized in that: In step 2, natural language processing is performed on the transaction data to parse out the transaction data of order information, delivery data and clearing and settlement data generated during the transaction process, and the transaction data is parsed to extract key information.

6. A dynamic compliance checking method for data transactions according to claim 1, characterized in that: In step 2, a multi-level data extraction method is used to parse and process the transaction data, specifically including: First, regular expressions are used to parse structured and semi-structured data to extract key fields of the transaction process. For unstructured text data, a word vector model is used to convert the text into a vector form to capture semantic information. Then, based on deep learning named entity recognition, key information related to the transaction process is identified; Dependency parsing is used to parse complex sentence structures and ensure the logical relationship of the identified key information in the context; Finally, the extracted key information is verified through the vector similarity algorithm.

7. An evaluation system for a dynamic compliance checking method for data transactions according to claim 1, comprising a data classification module, a data extraction module, a compliance policy generation module, and a policy evaluation and adjustment module, characterized in that: The data classification module collects transaction data of both parties to the transaction and classifies the transaction data according to the two parties to the transaction and the industry fields they belong to based on corresponding compliance rules; The data extraction module analyzes the transaction data and extracts key information related to the transaction process; The compliance strategy generation module inputs the classified transaction data and the key information related to the parsed transaction process into the dynamic compliance model, and generates a preliminary compliance strategy according to the preset compliance rules; The strategy evaluation and adjustment module dynamically analyzes historical transaction data and current transaction behavior based on the generated preliminary compliance strategy, evaluates the effectiveness of the compliance strategy and adjusts the compliance strategy.

8. The evaluation system according to claim 7, characterized in that: The strategy evaluation and adjustment module includes a transaction voucher generation module and a credit rating and compliance adjustment module; The transaction voucher generation module collects transaction process data in real time after each transaction is completed and generates a unique transaction voucher; The credit rating and compliance adjustment module generates an initial credit rating based on historical transaction records and transaction entity qualification information; dynamically adjusts the credit rating of both parties to the transaction based on the transaction voucher; Analyze each transaction behavior and its compliance results to obtain compliance trend forecasts for future transactions; Based on the compliance results and forecast results, adjust and analyze the current compliance strategy to evaluate whether the current strategy is suitable for future trends; Based on the analysis results, formulate a compliance strategy update plan and adjust the compliance strategies of both parties in future transactions accordingly.

Images