False trade identification method and system based on big data

By extracting and classifying key features of operating data in a big data environment, using machine learning models to identify false trade behaviors, solving the problem of low efficiency of traditional manual review, achieving efficient identification of false trade, reducing costs, and improving trade risk prevention Control ability.

CN119025889APending Publication Date: 2024-11-26CHINA ELECTRONICS TECHNOLOGY FINANCE CO LTD +1

Patent Information

Application Number
CN202411113879.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The traditional method of manually reviewing false trade is unable to efficiently identify false trade due to the large amount of information and large dimensions.

Method used

A false trade identification method based on big data is adopted to extract key elements of business data, extract trade data, extract key features, and use machine learning models to classify the identification trade behavior to generate identification results.

Benefits of technology

In the big data environment, greatly improve the identification efficiency of false trade, reduce labor costs, achieve rapid analysis of large-scale data, and improve trade risk prevention and control capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119025889A_ABST
    Figure CN119025889A_ABST
Patent Text Reader

Abstract

The invention discloses a big data-based false trade identification method and system, and belongs to the technical field of information identification. The method comprises the following steps: extracting key elements of operation data to obtain an information flow of the operation data; extracting trade data from the information flow of the operation data to generate a to-be-identified trade behavior; extracting key features of the trade behavior to be identified; and classifying a to-be-identified trade behavior based on the key feature to obtain an identification result of the to-be-identified trade behavior. According to the invention, the false trade identification efficiency is improved under big data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of information identification, and relates to a method and system for identifying false trade based on big data. Background Art

[0002] In recent years, all kinds of false trade have been strictly investigated, usually through manual verification. The audit department reviews the company's past transactions, contract signing and other information to identify whether the company has any violations of false trade regulations.

[0003] However, traditional manual review methods are costly, slow to respond, and highly dependent on human personal experience because they require verification of a large amount of information and multiple dimensions. Summary of the invention

[0004] In view of the above problems, the present invention discloses a method and system for identifying false trade based on big data, which adopts a model calculation method, thereby greatly improving the efficiency of identifying false trade in a big data environment and reducing labor costs.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution of the present invention includes the following contents.

[0006] A method for identifying false trade based on big data, the method comprising:

[0007] Extract key elements of business data to obtain the information flow of the business data; wherein the business data includes: sales contracts, transaction flow information, bill information and financing information;

[0008] Extracting trade data from the information flow of the business data to generate trade behaviors to be identified;

[0009] Extracting key features of the trade behavior to be identified;

[0010] The trade behavior to be identified is classified based on the key feature to obtain an identification result of the trade behavior to be identified.

[0011] Furthermore, the key elements of the sales contract include: contract type, counterparty name, contract signing date, contract amount, subject matter name and payment method; the key elements of the transaction flow information include: transaction type, counterparty name, payment date, payment amount and summary; the key elements of the bill information include: bill type, counterparty name, billing date and bill amount; the key elements of the financing information include: withdrawal object, withdrawal amount and withdrawal date.

[0012] Furthermore, trade data is extracted according to the information flow of the business data to generate trade behaviors to be identified, including:

[0013] Based on the information flow of the business data, obtaining the name of the subject matter in the sales contract, the sales contract including: a purchase contract and a sales contract;

[0014] Calculate the similarity of the names of the subject matter;

[0015] When the similarity between the names of the subject matter in the purchase contract and the sales contract is less than a first set threshold, a trade behavior to be identified is constituted based on the corresponding purchase contract and sales contract.

[0016] Furthermore, trade data is extracted according to the information flow of the business data to generate trade behaviors to be identified, including:

[0017] Based on the information flow of the business data, obtain the contract type and contract amount in the sales contract, the transaction type and payment amount in the transaction flow information, and the bill type and bill amount in the bill information;

[0018] According to the contract type in the sales contract, the transaction type in the transaction flow information, and the bill type in the bill information, the sales contract, the transaction flow information, and the bill information are classified into expenditure information and income information;

[0019] Calculate the difference rate between the expenditure amount and the income amount in the expenditure information and income information based on the contract amount in the sales contract, the payment amount in the transaction flow information and the bill amount in the bill information;

[0020] When the difference rate between the expenditure amount and the income amount is less than the second set threshold, a trade behavior to be identified is constituted based on the corresponding sales contract, transaction flow information and bill information.

[0021] Furthermore, trade data is extracted according to the information flow of the business data to generate trade behaviors to be identified, including:

[0022] Based on the information flow of the business data, obtain the contract type, counterparty name and contract amount in the sales contract, the transaction type, counterparty name and payment amount in the transaction flow information, and the bill type, counterparty name and bill amount in the bill information;

[0023] Obtain external enterprises based on the counterparty name in the sales contract, the counterparty name in the transaction flow information, and the counterparty name in the bill information;

[0024] For the sales contracts, transaction flow information and bill information involved in the same external enterprise, the sales contracts, transaction flow information and bill information are classified into expenditure information and income information according to the contract type in the sales contracts, the transaction type in the transaction flow information and the bill type in the bill information;

[0025] Calculate the difference rate of the total amount of revenue and expenditure of the external enterprise based on the contract amount in the sales contract, the payment amount in the transaction flow information, and the bill amount in the bill information;

[0026] When the difference rate of the total amount of income and expenditure of the external enterprise is less than the third set threshold, a trade behavior to be identified is constituted based on the corresponding sales contract, transaction flow information and bill information.

[0027] Furthermore, the key features of the trade behavior to be identified include: trade profit margin, whether the upstream and downstream are the same legal person, whether the upstream and downstream have the same shareholders, whether the upstream and downstream have cross-holdings, whether the upstream and downstream have industrial and commercial registration places, whether the telephone information is consistent, the presence of specific keywords in the sales contract, the loan behavior in the sales contract, whether the upstream and downstream have participated in the same bidding project, whether there have been disputes between the upstream and downstream, the nature of the upstream and downstream enterprises, whether the registration time of the upstream and downstream enterprises is less than y months, and whether the payment method of the purchase contract is better than the payment method of the sales contract.

[0028] Furthermore, the classification of the trade behavior to be identified based on the key feature to obtain the identification result of the trade behavior to be identified includes:

[0029] Use machine learning methods to build a fake trade identification model on training data;

[0030] The key features are classified based on the false trade identification model to obtain an identification result of the trade behavior to be identified.

[0031] Furthermore, the identification result of the trade behavior to be identified includes:

[0032] Inefficient trade based on trade profit margins;

[0033] and / or,

[0034] Trade in which there is a relationship between upstream and downstream, based on whether the upstream and downstream are the same legal person, whether the upstream and downstream are the same shareholders, whether the upstream and downstream have cross-holdings, and whether the upstream and downstream have the same business registration place and telephone information;

[0035] and / or,

[0036] Trades with financing risks based on the presence of specific keywords in the sales contract and the loan behavior in the sales contract;

[0037] and / or,

[0038] High-risk trades are obtained based on whether the upstream and downstream have participated in the same bidding project, whether there have been disputes between the upstream and downstream, the nature of the upstream and downstream enterprises, whether the registration time of the upstream and downstream enterprises is less than y months, and whether the payment method of the purchase contract is better than the payment method of the sales contract.

[0039] A false trade identification system based on big data, the system comprising:

[0040] An information extraction module is used to extract key elements of business data and obtain the information flow of the business data; wherein the business data includes: sales contracts, transaction flow information, bill information and financing information;

[0041] A behavior generation module is used to extract trade data according to the information flow of the business data to generate trade behaviors to be identified;

[0042] A feature extraction module, used to extract key features of the trade behavior to be identified;

[0043] The trade identification module is used to classify the trade behavior to be identified based on the key feature to obtain the identification result of the trade behavior to be identified.

[0044] An electronic device, characterized in that the electronic device comprises: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, it implements any of the above-mentioned methods for identifying false trade based on big data.

[0045] Compared with the prior art, the present invention first screens out trade behaviors in large-scale data through specific analysis of trade data, and further identifies risky and illegal trade based on specific technical characteristics of trade. Therefore, the present invention can quickly and frequently analyze large quantities of data. For enterprises with frequent business operations, the model capability of the present invention can replace the existing manual operation method to improve the ability to prevent and control trade risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Flowchart of the method for identifying corporate false trade based on big data. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in 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 work are within the scope of protection of the present invention.

[0048] The present invention integrates multi-dimensional big data, including internal financial capital flow data, bill data, financing data, contract data, and external industrial and commercial, judicial, and bidding data. Based on internal and external big data, the model intelligently analyzes and identifies trade behaviors, and then combines key information such as transaction time, amount, and counterparty to identify clues of suspected illegal trade.

[0049] like Figure 1As shown, the false trade identification method based on big data of the present invention comprises the following steps:

[0050] Step 1: Extract the key elements of the business data and obtain the information flow of the business data.

[0051] The information flow is composed of key elements extracted from the company's purchase and sale contracts (purchase contracts and sales contracts), transaction flow information, bill information, financing information and other data: including extracting fields from contract information: contract type, counterparty name, contract signing date, contract amount, subject matter name, payment method; extracting fields from transaction flow information: transaction type, counterparty name, payment date, payment amount, summary; extracting fields from bill information: bill type, counterparty name, invoicing date, bill amount; extracting fields from financing information: withdrawal object, withdrawal amount, withdrawal date.

[0052] Step 2: Extract trade data from the information flow of the business data to generate trade behaviors to be identified.

[0053] In one embodiment of the present invention, the trade behaviors to be identified can be divided into three categories.

[0054] Category 1: Identify the same subject matter of the transaction: For the contract information flow, search for purchase contracts and sales contracts signed by the enterprise within the same x months, and find purchase contracts and sales contracts with similar names of the subject matter of the transaction. If they exist, the identified contract information flow constitutes a trade.

[0055] Category 2: Identify transactions in similar time periods: Search for contracts / transactions / bills in the information flow with similar expenditure amounts and income amounts within the same x days, with a difference rate of less than n%. If such a difference exists, the identified information flow constitutes a trade.

[0056] Category 3: Identify the same buyer and seller: Check if the member units in the information flow have generated one or more income and expenditure contracts / transactions / bills with the same external enterprise within the past x months. If the difference rate of the total income and expenditure amount is less than n%, the identified information flow constitutes a trade.

[0057] Step 3: Extract key features of the trade behavior to be identified.

[0058] For a trade, the key elements in the information flow are extracted for preprocessing to obtain key features, which include: trade profit margin, whether the upstream and downstream are the same legal person, whether the upstream and downstream have the same shareholders, whether the upstream and downstream have cross-holdings, whether the upstream and downstream have industrial and commercial registration places, consistent telephone information, whether the upstream and downstream have participated in the same bidding project, whether there have been disputes between the upstream and downstream, the nature of the upstream and downstream enterprises, whether the registration time of the upstream and downstream enterprises is less than y months, and whether the payment method of the procurement contract is better than that of the sales contract.

[0059] Step 4: Classify the trade behavior to be identified based on the key feature to obtain the identification result of the trade behavior to be identified.

[0060] The extracted key features of trade are put into the classification model to mark the trade. Inefficient business, idle trade, and suspected financing trade are identified. The classification model can be a machine learning model or a deep learning model, which is not limited by the present invention.

[0061] The recognition results can be mainly divided into the following four categories:

[0062] Inefficient trade: For the identified trade, analyze the trade profit margin to see whether the buying and selling amounts are similar or even almost the same. If the profit margin is less than n%, it is identified as an inefficient trade business.

[0063] Trade with associated relationships between upstream and downstream companies: For upstream and downstream manufacturers involved in trade, use industrial and commercial information, bidding information, and judicial information to determine whether there is a special interest relationship. For example, if the legal person of the upstream and downstream companies is the same, the upstream and downstream companies have the same shareholders, the upstream and downstream companies cross-hold shares, and the upstream and downstream companies have the same industrial and commercial addresses. If any of the above characteristics are met, it is determined that the upstream and downstream companies are associated.

[0064] Trade with financing risks: Search for keywords such as financing in the contract terms; identify whether any member unit has taken out a bank loan within x days before making an outbound remittance during the fund transaction, and whether the difference between the loan amount M and the external expenditure amount N is less than Q%. If so, identify the risk of financing trade behavior.

[0065] High-risk trade: if the upstream and downstream enterprises have participated in the same bidding projects many times, there have been disputes between the upstream and downstream enterprises, the upstream and downstream enterprises are private enterprises, the registration time of the upstream and downstream enterprises is less than y months, the payment method of the purchase contract is a one-time payment and the sales contract is an installment payment, the more key features are hit, the higher the total risk weight value.

[0066] The system is regularly updated through the model, and the recognition thresholds for time, amount, etc. can be customized. After the model identifies illegal trade, the statistics are visualized in the system, and the recognition results and problem details are displayed.

[0067] In summary, the illegal trade identification method adopted by the present invention uses data of various dimensions and covers a wide range, identifying trade behaviors from both the financial and contractual aspects. The internal enterprise data used includes transaction flows, bills, and financing. In addition, it integrates external industrial and commercial, judicial, bidding and other multi-dimensional data for analysis and judgment, and effectively mines the characteristics and clues of illegal trade. At the same time, the scope of trade identification is also analyzed in this method in multiple dimensions, analyzing the behaviors of member units themselves conducting trade with external manufacturers, and the behaviors of directly affiliated units conducting trade with external manufacturers.

[0068] The present invention uses industrial and commercial, judicial, and bidding data to directly and indirectly analyze whether there is an association relationship between upstream and downstream trade enterprises. Industrial and commercial data identifies the association relationship by analyzing whether the enterprises have the same legal person, the same registered phone number, the same registered address, the same shareholders, and cross-holdings; at the same time, based on the bidding data, by analyzing whether the enterprises have participated in the same bidding project, and the judicial data, by analyzing whether the enterprises have had disputes, it is analyzed from the side whether the enterprises have had associations in the past, and the trade risk coefficient is weighted to increase.

[0069] The above implementation is only used to illustrate the technical solution of the present invention rather than to limit it. Ordinary technicians in this field can modify or replace the technical solution of the present invention with equivalents without departing from the scope of the present invention. The protection scope of the present invention shall be based on the claims.

Claims

1. A method for identifying fake trade based on big data, It is characterized in that The method comprises: Extract key elements of business data to obtain the information flow of the business data; wherein the business data includes: sales contracts, transaction flow information, bill information and financing information; Extracting trade data from the information flow of the business data to generate trade behaviors to be identified; Extracting key features of the trade behavior to be identified; The trade behavior to be identified is classified based on the key feature to obtain an identification result of the trade behavior to be identified.

2. The method according to claim 1, It is characterized in that The key elements of the sales contract include: contract type, counterparty name, contract signing date, contract amount, subject matter name and payment method; the key elements of the transaction flow information include: transaction type, counterparty name, payment date, payment amount and summary; the key elements of the bill information include: bill type, counterparty name, billing date and bill amount; the key elements of the financing information include: withdrawal object, withdrawal amount and withdrawal date.

3. The method according to claim 2, It is characterized in that Extract trade data based on the information flow of the business data to generate trade behaviors to be identified, including: Based on the information flow of the business data, obtaining the name of the subject matter in the sales contract, the sales contract including: a purchase contract and a sales contract; Calculate the similarity of the names of the subject matter; When the similarity between the names of the subject matter in the purchase contract and the sales contract is less than a first set threshold, a trade behavior to be identified is constituted based on the corresponding purchase contract and sales contract.

4. The method according to claim 2, It is characterized in that Extract trade data based on the information flow of the business data to generate trade behaviors to be identified, including: Based on the information flow of the business data, obtain the contract type and contract amount in the sales contract, the transaction type and payment amount in the transaction flow information, and the bill type and bill amount in the bill information; According to the contract type in the sales contract, the transaction type in the transaction flow information, and the bill type in the bill information, the sales contract, the transaction flow information, and the bill information are classified into expenditure information and income information; Calculate the difference rate between the expenditure amount and the income amount in the expenditure information and income information based on the contract amount in the sales contract, the payment amount in the transaction flow information and the bill amount in the bill information; When the difference rate between the expenditure amount and the income amount is less than the second set threshold, a trade behavior to be identified is constituted based on the corresponding sales contract, transaction flow information and bill information.

5. The method according to claim 2, It is characterized in that Extract trade data based on the information flow of the business data to generate trade behaviors to be identified, including: Based on the information flow of the business data, obtain the contract type, counterparty name and contract amount in the sales contract, the transaction type, counterparty name and payment amount in the transaction flow information, and the bill type, counterparty name and bill amount in the bill information; Obtain external enterprises based on the counterparty name in the sales contract, the counterparty name in the transaction flow information, and the counterparty name in the bill information; For the sales contracts, transaction flow information and bill information involved in the same external enterprise, the sales contracts, transaction flow information and bill information are classified into expenditure information and income information according to the contract type in the sales contracts, the transaction type in the transaction flow information and the bill type in the bill information; Calculate the difference rate of the total amount of revenue and expenditure of the external enterprise based on the contract amount in the sales contract, the payment amount in the transaction flow information, and the bill amount in the bill information; When the difference rate of the total amount of income and expenditure of the external enterprise is less than the third set threshold, a trade behavior to be identified is constituted based on the corresponding sales contract, transaction flow information and bill information.

6. The method according to claim 2, It is characterized in that The key features of the trade behavior to be identified include: trade profit margin, whether the upstream and downstream are the same legal person, whether the upstream and downstream have the same shareholders, whether the upstream and downstream have cross-holdings, whether the upstream and downstream have industrial and commercial registration places, whether the telephone information is consistent, the presence of specific keywords in the sales contract, the loan behavior in the sales contract, whether the upstream and downstream have participated in the same bidding project, whether there have been disputes between the upstream and downstream, the nature of the upstream and downstream enterprises, whether the registration time of the upstream and downstream enterprises is less than y months, and whether the payment method of the purchase contract is better than the payment method of the sales contract.

7. The method according to claim 6, It is characterized in that The classifying the trade behavior to be identified based on the key feature to obtain the identification result of the trade behavior to be identified includes: Use machine learning methods to build a fake trade identification model on training data; The key features are classified based on the false trade identification model to obtain an identification result of the trade behavior to be identified.

8. The method according to claim 7, It is characterized in that The identification result of the trade behavior to be identified includes: Inefficient trade based on trade profit margins; and / or, Trade in which there is a relationship between upstream and downstream, based on whether the upstream and downstream are the same legal person, whether the upstream and downstream are the same shareholders, whether the upstream and downstream have cross-holdings, and whether the upstream and downstream have the same business registration place and telephone information; and / or, Trades with financing risks based on the presence of specific keywords in the sales contract and the loan behavior in the sales contract; and / or, High-risk trades are obtained based on whether the upstream and downstream have participated in the same bidding project, whether there have been disputes between the upstream and downstream, the nature of the upstream and downstream enterprises, whether the registration time of the upstream and downstream enterprises is less than y months, and whether the payment method of the purchase contract is better than the payment method of the sales contract.

9. A fake trade identification system based on big data, It is characterized in that The system comprises: An information extraction module is used to extract key elements of business data and obtain the information flow of the business data; wherein the business data includes: sales contracts, transaction flow information, bill information and financing information; A behavior generation module is used to extract trade data according to the information flow of the business data to generate trade behaviors to be identified; A feature extraction module, used to extract key features of the trade behavior to be identified; The trade identification module is used to classify the trade behavior to be identified based on the key feature to obtain the identification result of the trade behavior to be identified.

10. An electronic device, It is characterized in that The electronic device comprises: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the method for identifying false trade based on big data as described in any one of claims 1 to 8 is implemented.

Citation Information

Patent Citations

  • Big data intelligent marketing system and method thereof

    CN110119466A

  • Enterprise financing trade analysis method and device

    CN110955735A

  • Malicious associated transaction monitoring method and device based on blockchain and computer equipment

    CN111861476A

  • Enterprise trade information visual display method based on knowledge graph

    CN116541537A

Cited By

  • Building industry false trade early warning method and device

    CN119692608A