CN113011973A - Financial transaction supervision model, system and equipment based on intelligent contract data lake - Google Patents
Financial transaction supervision model, system and equipment based on intelligent contract data lake Download PDFInfo
- Publication number
- CN113011973A CN113011973A CN202110084721.5A CN202110084721A CN113011973A CN 113011973 A CN113011973 A CN 113011973A CN 202110084721 A CN202110084721 A CN 202110084721A CN 113011973 A CN113011973 A CN 113011973A
- Authority
- CN
- China
- Prior art keywords
- data
- transaction
- result
- database
- supervision
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The application discloses a financial transaction supervision model, a system and equipment based on an intelligent contract data lake. The specific process is as follows: acquiring transaction data; the prediction machine judges whether the prediction machine accords with the uplink condition; the transaction data are judged to be in compliance by the prediction machine and then transmitted into an intelligent contract data lake, wherein the intelligent contract data lake consists of a MySQL database, a cache database and an intelligent contract database, the MySQL database stores the transaction data, the cache database stores high-frequency high-risk transaction data, and the intelligent contract database stores supervision flow data; the transaction data is directed by the intelligent controller to sequentially execute operations in an early supervision stage, a middle supervision stage and a later supervision stage in a machine learning engine; the machine learning engine combines a KYC (user identity recognition), AML (anti-money laundering) method and a machine learning algorithm based on rules to classify the risk of the transaction, so that the functions of link analysis, behavior modeling, risk early warning, anomaly detection and transaction grading are realized, and a final judgment result is obtained; the intelligent controller executes a transaction passing or transaction withdrawing operation according to the result.
Description
Technical Field
The invention belongs to the field of block chains, and particularly relates to a financial transaction supervision model, system and equipment based on an intelligent contract data lake.
Background
On the second outband financial peak just ended, the concepts of "digital economy", "cloud and chain", "financial technology", "regulatory framework", etc. are repeatedly mentioned. As an important content of a modern financial supervision system, anti-money laundering is an important guarantee for maintaining economic and social stability, is an important hand for practically preventing financial risks and optimizing industry supervision effects, and is also an important means for participating in global management and expanding the bidirectional opening of the financial industry. At present, the financial industry forms a relatively system-complete financial industry anti-money laundering system which takes ' anti-money laundering law ' as a basic law, takes rules of anti-money laundering related departments established by people's banks, silver insurance guidances and certificate guidances as specific guidance, and takes self-discipline issued by industry associations as specific implementation. However, the anti-money laundering supervision still has a short board, the mechanism of the system needs to be further improved, the supervision effectiveness needs to be further enhanced, and the supervision means needs to be further improved.
A common financial regulatory platform consists of two parts: a customer identity identification module and an anti-money laundering module. The identity recognition of the client is carried out by the methods of biological information recognition such as the due-employment investigation, the enhanced due-employment investigation, SWIFT filtering, fingerprint face and the like of the client. This step requires a large number of documents, spreadsheets and other tools to compare the records of the customer information, which can consume labor and time costs, reduce customer stickiness and transaction efficiency, and cause privacy problems in some countries and regions. While the anti-money laundering module includes conventional rule engine-based detection methods and detection schemes using machine learning techniques. The traditional detection method based on the rule engine cannot identify suspicious behaviors in low-frequency transfer transactions and complex transaction money laundering behaviors in massive transactions. The machine learning algorithm has low diagnosis accuracy on complex associated transactions and sporadic low-frequency transactions.
In 2019, the blockchain is taken as a core technology in China to independently innovate important breakthrough, and the innovation development of the blockchain technology and industry is promoted quickly. The traditional financial industry supervision institutions pay huge supervision cost on the anti-money laundering work, and obtain certain results on the anti-money laundering work, but from the current anti-money laundering mechanism of each financial institution, the problems of low customer identity recognition efficiency, low information degree of the anti-money laundering work, high anti-money laundering supervision cost, asynchrony and non-sharing of related data among the financial institutions and the like still exist. Therefore, the block chain is introduced to realize the innovation of intra-department supervision, the block chain technology is introduced into the links of daily customer identity identification, anti-money laundering detection, transaction audit and the like of a financial institution, and the digital, automatic and intelligent real-time supervision of supervision rules before, during and after transaction is realized by using an intelligent contract.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a financial transaction supervision model, system and equipment based on an intelligent contract data lake, so that the supervision accuracy and efficiency are improved, and the manpower resource and time cost are reduced.
One aspect of the invention provides a financial transaction supervision model based on an intelligent contract data lake, which comprises the following steps:
step S10, a machine learning data set with financial characteristic attributes is arranged from a UCI database, and a training data set and a testing data set required by an experiment are obtained after data preprocessing;
step S20, the experimental data set is used as a data source to send contract calling to an Oraclize prespecification machine, the Oraclize prespecification machine inquires and checks the compliance and then guides the experimental data set into an intelligent contract data lake, and if the rule is not matched, the transaction is terminated;
and step S30, the experimental data are respectively stored in a MySQL database, a Cache database and a Smart Contract database according to factors such as attributes, characteristics, types and processing stages. The MySQL database stores all data types of experimental data before the experimental data are subjected to supervision operation, and the Cache database stores data types of short-interval high-frequency fine-grained calling, such as: relationship data, account data, tax data, historical data, scoring data, blacklist \ white list data, and the like. The Smart Contract database stores the data type of the transaction characteristic attribute;
step S40, under the command of the intelligent controller, the characteristic data of different areas on the intelligent contract data are sequentially transmitted to the machine learning engine to execute the operations of early supervision, middle supervision and later supervision, which respectively correspond to the following steps: KYC (customer identification), AML (anti-money laundering test), Credit grading. The intelligent controller firstly calls account data, score data, blacklist/white list and other data cached in a Cache database, and performs KYC operation on the data. If the execution result passes, step S50 is entered, otherwise, the transaction is terminated;
and step S50, under the direction of the intelligent controller, the characteristic attribute data in the Smart Contract database is executed with AML operation, and the transaction data determines the execution result after three methods of behavior modeling, link analysis and anomaly detection arbitration decision. If the execution result passes, step S60 is entered, otherwise, the transaction is terminated;
and step S60, under the direction of the intelligent controller, the data passed in the step S50 and the characteristic attribute data mapped in the Smart Contract database are subjected to Credit grading operation, and the transaction data obtains corresponding scores in the evaluation of the grading card model to be used as the Credit score of the transaction. The intelligent controller stores the scoring result in a Cache database;
and step S70, the intelligent controller returns the final judgment result of the machine learning engine to the Smart Contract database to determine the final transaction result and display the transaction condition and the prediction accuracy.
In some preferred embodiments, the data preprocessing method in step S10 is:
and checking data by adopting a head () method, processing missing data, correspondingly adding default values, deleting incomplete rows and columns, and storing results after normalizing data types.
In some preferred embodiments, in step S20, the Oraclize presiding machine calls different layers from bottom to top in sequence to perform the ping operation, and the logical structure is:
the network protocol is a network topology structure of a centralized prediction machine, and a single centralized service provider controls an intermediate node;
the execution of smart contracts and data calls in the operational layer are both performed on Trusted Execution Environments (TEEs). The AWS takes the audit role and integrity is verified by tlsnottary Proof. Relying on a multi-signature mechanism to allow speakers (Oracles) meeting more than the minimum number of honest nodes to simultaneously sign corresponding nodes;
the contract layer includes order matching contracts, service request contracts, data invocation interfaces, and service standard protocols.
In some preferred embodiments, in step S40, "the intelligent controller first retrieves account data, score data, and data such as a black list \ white list cached in the Cache database, and performs KYC operation on the account data, score data, and data such as a black list \ white list," and the method includes:
step S401, performing Digital organizing and SWIFT filtering operation on the input data, if the result is positive, entering step S402, and if the result is negative, returning, namely terminating the transaction;
step S402, carrying out CDD (customer due diligence) and EDD (enhanced due diligence investigation) operation on the incoming data, if the result is positive, entering step S403, and if the result is negative, returning, namely terminating the transaction;
in step S403, Whitelist/Blacklist Filter operation is performed on the incoming data, and if the result passes, step S50 is performed, otherwise, the transaction is returned, i.e., terminated.
In some preferred embodiments, step S50 "the transaction data determines the execution result after arbitration decision of triple method of behavior modeling, link analysis and anomaly detection", which is performed by:
step S501, an SVM (Support Vector Machine) algorithm is adopted to perform behavior modeling three-classification operation on the transaction data, and the results are respectively: secure transactions, suspicious transactions, pending transactions. If the result is a secure transaction, go to step S503; if the result is suspicious transaction, returning, namely terminating the transaction; if the result is a pending transaction, the step S502 is executed; the SVM algorithm selects a Sigmoid kernel function, and the calculation method comprises the following steps:
wherein X1，X2Is data corresponding to two classes, κ (X)1,X2) Is a sufficient condition for positive definite core, a is used to set gamma parameter setting in kernel function, default is 1/k (k is number of classes), b is used to set coef0 in kernel function, default is 0;
step S502, a MaxEnt algorithm (Max entry, maximum Entropy) is adopted for performing link analysis operation on the transaction to be determined, and if the classification result is safe transaction, the step S503 is performed; if the result is suspicious transaction, returning, namely terminating the transaction;
step S503, adopt
The Bayesian algorithm detects the abnormality of the secure transaction, if the result is positive, the process goes to step S60, and if the result is negative, the process returns, i.e., the transaction is terminated.
In some preferred embodiments, step S60 "transaction data gets corresponding score in the evaluation of the score card model" is performed by:
step S601, a bucket dividing method is adopted, each processing value is endowed with a corresponding attribute, and numerical value characteristics are converted into classification characteristics;
step S602, calculating an evidence weight (WoE) of each attribute and an Information Value (IV) of each feature point, where the calculation formula of the evidence weight is:
[ln(Distr G/Distr B)]×100
wherein, G represents that the customer transaction passes the target variable being 0, B represents that the customer transaction refuses the target variable being 1;
the information value calculation formula is as follows:
wherein, G represents that the customer transaction passes the target variable being 0, B represents that the customer transaction refuses the target variable being 1;
step S603, model building is performed by replacing the value of the original variable with WoE, and model selection LR (Logistic regression) has the expression:
where y is the probability that the label is A, x is the predicted label, w is the training parameter, w is the probability that the label is ATIs a weight value;
step S604, adopting cross validation and grid search to adjust parameters, converting the parameters into two classification problems, wherein the loss function is as follows:
wherein F (w) is a loss function, n is a sample number, ynIs a sample label, p is the corresponding probability;
step S605, calculating a score coefficient of the score card for each attribute to obtain a final score card. The scoring formula is:
where β is an LR coefficient of a given attribute, α is an LR intercept, WoE is an evidence weight of the given attribute, n is a model feature number, Factor, and Offset are scaling parameters.
In some preferred embodiments, the intelligent contract processing method of step S70 is:
step S701, checking whether the information submitted when the user is established and the transaction amount are true and legal;
step S702, checking whether the transfer initiator and the beneficiary are legal users;
and step S703, judging whether the contract operation is continuously executed according to the result returned by the machine learning engine.
In some preferred embodiments, the financial transaction regulatory model prediction accuracy information may be presented in a bar graph.
In another aspect, the present invention provides a financial transaction monitoring system based on intelligent contract data lake, the system comprising: the system comprises a data deep processing module, a feature marking module, a prediction machine module, an intelligent contract data lake module, a machine learning engine module, an intelligent controller module and an accuracy rate display module;
the data deep processing module is configured to carry out data cleaning data preprocessing operation on the machine learning data set sorted out from the UCI database, and store the result as a transaction data set;
the data feature marking module is configured to construct a six-dimensional feature data set from the transaction data set;
the Oraclize presbyope module is configured to query and check the transaction data through the Oraclize presbyope module, determine whether the transaction data is in compliance, and then execute the next operation;
the intelligent Contract data lake module is configured to store the transaction data in a MySQL database, a Cache database and a Smart Contract database respectively according to different transaction data attributes, characteristics, types and processing stages;
the machine learning engine module is configured to sequentially transmit the feature data of different regions into the machine learning engine to execute early supervision, middle supervision and later supervision operations;
the intelligent controller module is configured to enable the intelligent controller to have the function of unified command data, algorithm, block and database combined operation;
and the accuracy rate display module is configured to display a final transaction structure, a transaction condition and a prediction accuracy rate.
In a third aspect of the invention, a storage device is provided, in which a program is stored, the program being adapted to be loaded and executed by a processor to implement a smart contract data lake based financial transaction supervision model as described above.
In a fourth aspect of the present invention, a processing apparatus is provided, comprising: a processor and a memory;
the processor is adapted to execute a program, and the memory is adapted to store the program;
the program is adapted to be loaded and executed by the processor to implement a smart contract data lake based financial transaction regulatory model as described above.
The invention has the beneficial effects that:
the financial transaction supervision model based on the intelligent contract data lake is disclosed by the invention. The problems that the existing model algorithm is low in identification precision of small-amount high-frequency transactions, complex associated transactions and accidental low-frequency transactions and the historical transactions are difficult to trace are solved; the money laundering behavior of mass transactions and complex transaction means can be effectively identified; the limitations of algorithm singleness, inexplicability and the like are improved; the accuracy and the efficiency of prediction are improved, the labor cost is reduced, and the pressure of the server is relieved.
Drawings
FIG. 1 is a schematic flow diagram of the financial transaction supervisory model of the present invention based on the intelligent contract data lake;
FIG. 2 is a system block diagram of the financial transaction supervisory model of the present invention based on the intelligent contract data lake;
FIG. 3 is an organizational architecture diagram of the financial transaction supervisory model of the present invention based on the Intelligent contract data lake;
FIG. 4 is a diagrammatic view of an intelligent contract data lake embodiment of the financial transaction supervisory model of the present invention based on the intelligent contract data lake;
FIG. 5 is a logic diagram of a machine learning system of an embodiment of the financial transaction supervisory model based on the intelligent contract data lake of the present invention;
FIG. 6 is a precision diagram of the predicted results of an embodiment of the financial transaction supervisory model based on the intelligent contract data lake of the present invention.
Detailed description of the preferred embodiments
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the related invention are shown in the drawings.
It should be noted that the embodiments and features in the embodiments in the present application may be combined with each other without conflict.
The invention relates to a financial transaction supervision model based on an intelligent contract data lake. The problems that the existing model algorithm is low in identification precision of small-amount high-frequency transactions, complex associated transactions and accidental low-frequency transactions and the historical transactions are difficult to trace are solved; the money laundering behavior of mass transactions and complex transaction means can be effectively identified; the limitations of single and unexplainable methods are improved; the accuracy and the efficiency of prediction are improved, the labor cost is reduced, and the pressure of the server is relieved.
For a clear explanation of the financial transaction supervision model based on the intelligent contract data lake, the following is a detailed description of the steps in the method embodiment of the present invention with reference to fig. 1.
Step S10, a machine learning data set with financial characteristic attributes is arranged from a UCI database, and a training data set and a testing data set required by an experiment are obtained after data preprocessing;
using a database to store a machine learning data set with financial characteristic attributes, which is arranged in a UCI database, wherein the data set needs to contain six types of characteristics, which are respectively: basic information, customer portrait, account dimension, transaction amount, transaction stroke dimension, and opponent dimension.
In the preferred embodiment of the invention, the Python is used for realizing the cleaning pretreatment work of the data, the head () method is adopted to check the data and process the missing data, the default value is correspondingly added, then the incomplete row and column are deleted, the data type is normalized and the missing item is complemented to obtain the transaction data.
Step S20, the experimental data set is used as a data source to send contract calling to an Oraclize prespecification machine, the Oraclize prespecification machine inquires and checks the compliance and then guides the experimental data set into an intelligent contract data lake, and if the rule is not matched, the transaction is terminated;
in the preferred embodiment of the invention, the Oraclize prescaler calls different layers from bottom to top in sequence to execute the inspection operation on the transaction data, and the logic structure is as follows:
the network protocol is configured such that a single centralized service provider controls a single intermediate node; the execution of smart contracts and data calls in the operational layer are both performed on Trusted Execution Environments (TEEs). The AWS takes the audit role and integrity is verified by tlsnottary Proof. Relying on a multi-signature mechanism to allow speakers (Oracles) meeting more than a minimum number of honest nodes to simultaneously sign transaction data; the contract layer determines whether the service request contract passes.
And step S30, the experimental data are respectively stored in a MySQL database, a Cache database and a Smart Contract database according to factors such as attributes, characteristics, types and processing stages. The MySQL database stores all data types of experimental data before the experimental data are subjected to supervision operation, and the Cache database stores data types of short-interval high-frequency fine-grained calling, such as: relationship data, account data, tax data, historical data, scoring data, blacklist \ white list data, and the like. The Smart Contract database stores the data type of the transaction characteristic attribute;
six-dimensional feature data in the transaction data are stored in a MySQL database; storing information such as the types, attributions, offshore accounts, high-risk areas, account opening duration and the like of the customers in a Cache database; and storing information such as multi-currency transactions, rapid increase of cash withdrawal, rapid increase of large-amount consumption, loan proportion, small-amount transfer statistical characteristics and the like related to the transaction characteristics into a Smart Contract database so as to facilitate the acquisition and calling of the intelligent controller. The system structure diagram is shown in fig. 2, and the detailed architecture diagram is shown in fig. 3.
Step S40, under the command of the intelligent controller, the characteristic data of different areas on the intelligent contract data are sequentially transmitted to the machine learning engine to execute the operations of early supervision, middle supervision and later supervision, which respectively correspond to the following steps: KYC (customer identification), AML (anti-money laundering test), Credit grading. The intelligent controller firstly calls account data, score data, blacklist/white list and other data cached in a Cache database, and performs KYC operation on the data. If the execution result passes, step S50 is entered, otherwise, the transaction is terminated;
in the step, the transaction data is under the command of the intelligent controller, and operations of KYC (corresponding to early supervision), AML (corresponding to middle supervision) and Credit grading (corresponding to later supervision) are sequentially executed in the machine learning engine, and the trend of the next step is determined according to the result. The specific operation principle can refer to fig. 4.
In a preferred embodiment of the present invention, step S40, "the intelligent controller first retrieves account data, score data, and data such as a black list \ white list cached in the Cache database, and performs KYC operation on the data," the method includes:
step S401, performing Digital organizing and SWIFT filtering operation on the input data, if the result is positive, entering step S402, and if the result is negative, returning, namely terminating the transaction;
in this step, incoming data first passes through basic information in the model, such as: and performing primary digital enrollment operation on the information such as name, age, place of birth, transaction area, beneficiary account and the like, and then further verifying whether the identity of the client is in compliance by using a biological information identification mode such as face identification and the like.
Step S402, carrying out CDD (customer due diligence) and EDD (enhanced due diligence investigation) operation on the incoming data, if the result is positive, entering step S403, and if the result is negative, returning, namely terminating the transaction;
in this step, attributes in the incoming data, such as location of ownership, age, bank clerk, money laundering risk, multi-currency transactions, etc., perform CDD and EDD operations in the model.
In step S403, Whitelist/Blacklist Filter operation is performed on the incoming data, and if the result passes, step S50 is performed, otherwise, the transaction is returned, i.e., terminated.
The system scores different customers in past historical transactions, and obtains a white list and a black list according to the scores of the customers. This step quickly determines whether the transacting client is in compliance by way of a query.
And step S50, under the direction of the intelligent controller, the characteristic attribute data in the Smart Contract database is executed with AML operation, and the transaction data determines the execution result after three methods of behavior modeling, link analysis and anomaly detection arbitration decision. If the execution result passes, step S60 is entered, otherwise, the transaction is terminated;
in the step, an arbitration model combining triple detection methods of behavior modeling, link analysis and anomaly detection is designed, 3 methods jointly vote to determine whether the transaction finally passes, and if any method in the arbitration model applies negative vote, the transaction is terminated. A machine learning engine is formed by combining a KYC model in the early stage of supervision and a scoring model in the later stage of supervision, and a logic diagram of the engine is shown in FIG. 5.
In a preferred embodiment of the present invention, step S50 "determining an execution result after arbitration decision of triple methods including behavior modeling, link analysis, and anomaly detection" includes:
step S501, an SVM (Support Vector Machine) algorithm is adopted to perform behavior modeling three-classification operation on the transaction data, and the results are respectively: secure transactions, suspicious transactions, pending transactions. If the result is a secure transaction, go to step S503; if the result is suspicious transaction, returning, namely terminating the transaction; if the result is a pending transaction, the step S502 is executed; the SVM algorithm selects a Sigmoid kernel function, and the calculation method comprises the following steps:
wherein X1，X2Is data corresponding to two classes, κ (X)1,X2) Is a sufficient condition for positive definite core, a is used to set gamma parameter setting in kernel function, default is 1/k (k is number of classes), b is used to set coef0 in kernel function, default is 0;
step S502, a MaxEnt algorithm (Max entry, maximum Entropy) is adopted for performing link analysis operation on the transaction to be determined, and if the classification result is safe transaction, the step S503 is performed; if the result is suspicious transaction, returning, namely terminating the transaction;
step S503, adopt
The Bayesian algorithm detects the abnormality of the secure transaction, if the result is positive, the process goes to step S60, and if the result is negative, the process returns, i.e., the transaction is terminated.
Taking a certain number of transactions in the test data set as an example, the transactions numbered 0007, 0005 and 0524 are all judged to be safe transactions, and directly enter a scoring module in the later supervision stage. The transaction numbered 0217 and 0479 is judged to be suspicious due to the fact that the transaction is over-regional, over-currency, account opening time period and age of the client is too big, the transaction is rejected, and the result is directly returned to Smart Contract data lake.
And step S60, under the direction of the intelligent controller, the data passed in the step S50 and the characteristic attribute data mapped in the Smart Contract database are subjected to Credit grading operation, and the transaction data obtains corresponding scores in the evaluation of the grading card model to be used as the Credit score of the transaction. The intelligent controller stores the scoring result in a Cache database;
in a preferred embodiment of the present invention, step S60 "the transaction data gets a corresponding score in the evaluation of the score card model", the method comprises:
step S601, a bucket dividing method is adopted, each processing value is endowed with a corresponding attribute, and numerical value characteristics are converted into classification characteristics;
step S602, calculating an evidence weight (WoE) of each attribute and an Information Value (IV) of each feature point, where the calculation formula of the evidence weight is:
[ ln (Distr G/Distr B) ]. times.100 formula (2)
Wherein, G represents that the customer transaction passes the target variable being 0, B represents that the customer transaction refuses the target variable being 1;
the information value calculation formula is as follows:
wherein, G represents that the customer transaction passes the target variable being 0, B represents that the customer transaction refuses the target variable being 1;
step S603, model building is performed by replacing the value of the original variable with WoE, and model selection LR (Logistic regression) has the expression:
where y is the probability that the label is A, x is the predicted label, w is the training parameter, w is the probability that the label is ATIs a weight value;
step S604, adopting cross validation and grid search to adjust parameters, converting the parameters into two classification problems, wherein the loss function is as follows:
wherein F (w) is a loss function, n is a sample number, ynIs a sample label, p is the corresponding probability;
step S605, calculating a score coefficient of the score card for each attribute to obtain a final score card. The scoring formula is:
where β is an LR coefficient of a given attribute, α is an LR intercept, WoE is an evidence weight of the given attribute, n is a model feature number, Factor, and Offset are scaling parameters.
Taking the determined result as a certain number of transactions of the safe transactions as an example, three primary variables are set, the first variable is a target variable, the binary classification variable is adopted, and the rest variables are characteristics. In the link of feature prediction, 8 features are selected according to the IV value for model training, and Factor is 28.85 and Offset is 487.14 after calculation. A determination of the final score is made, for example: one trader is 45 years old, the liability rate is 0.5, and the monthly income is 50000 RMB. It scores 53+55+ 57-165 and may be white listed. Traders with 2 suspicious transactions within 3 years are blacklisted.
And step S70, the intelligent controller returns the final judgment result of the machine learning engine to the Smart Contract database to determine the final transaction result and display the transaction condition and the prediction accuracy.
In a preferred embodiment of the present invention, the intelligent contract processing method in step S70 includes:
step S701, checking whether the information submitted when the user is established and the transaction amount are true and legal;
step S702, checking whether the transfer initiator and the beneficiary are legal users;
and step S703, judging whether the contract operation is continuously executed according to the result returned by the machine learning engine.
In this step, if the transaction type is a safe transaction, the above operations are not executed manually, and if the transaction type is a suspicious transaction, manual verification is required to determine whether the transaction on the intelligent contract passes or not. Transactions numbered 0007, 0005, 0524 are passed. The transactions numbered 0217 and 0479 all need manual secondary judgment.
In a preferred embodiment of the invention, the information of the prediction accuracy of the financial transaction supervision model can be displayed by a bar chart.
As shown in fig. 6, a bar chart showing the accuracy of the transaction managed model to determine the correct transaction according to the time variation in this experiment is shown. The prediction accuracy rate is not changed along with time, but is only related to the corresponding parameters, and the prediction accuracy rate of the algorithm can be effectively increased by adopting different parameters corresponding to different types of transactions.
The financial transaction supervision system based on the intelligent contract data lake comprises a data deep processing module, a feature marking module, a prediction machine module, an intelligent contract data lake module, a machine learning engine module, an intelligent controller module and an accuracy rate display module;
the data deep processing module is configured to carry out data cleaning data preprocessing operation on the machine learning data set sorted out from the UCI database, and store the result as a transaction data set;
the data feature marking module is configured to construct a six-dimensional feature data set from the transaction data set;
the Oraclize presbyope module is configured to query and check the transaction data through the Oraclize presbyope module, determine whether the transaction data is in compliance, and then execute the next operation;
the intelligent Contract data lake module is configured to store the transaction data in a MySQL database, a Cache database and a Smart Contract database respectively according to different transaction data attributes, characteristics, types and processing stages;
the machine learning engine module is configured to sequentially transmit the feature data of different regions into the machine learning engine to execute early supervision, middle supervision and later supervision operations;
the intelligent controller module is configured to enable the intelligent controller to have the function of unified command data, algorithm, block and database combined operation;
and the accuracy rate display module is configured to display a final transaction structure, a transaction condition and a prediction accuracy rate.
A storage device according to a third embodiment of the present invention stores therein a program adapted to be loaded and executed by a processor to implement a smart contract data lake-based financial transaction monitoring model as described above.
A processing apparatus according to a fourth embodiment of the present invention includes: a processor and a memory; the processor is adapted to execute a program, and the memory is adapted to store the program; the program is adapted to be loaded and executed by the processor to implement a smart contract data lake based financial transaction regulatory model as described above.
Those of skill in the art would appreciate that the various illustrative modules, method steps, and modules described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that programs corresponding to the software modules, method steps may be located in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. To clearly illustrate this interchangeability of electronic hardware and software, various illustrative components and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as electronic hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The terms "comprises," "comprising," or any other similar term are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
So far, the technical solutions of the present invention have been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of the present invention is obviously not limited to these specific embodiments. Equivalent changes or substitutions of related technical features can be made by those skilled in the art without departing from the principle of the invention, and the technical scheme after the changes or substitutions can fall into the protection scope of the invention.
Claims (11)
1. A financial transaction supervisory model based on a smart contract data lake, the model comprising:
step S10, a machine learning data set with financial characteristic attributes is arranged from a UCI database, and a training data set and a testing data set required by an experiment are obtained after data preprocessing;
step S20, the experimental data set is used as a data source to send contract calling to an Oraclize prespecification machine, the Oraclize prespecification machine inquires and checks the compliance and then guides the experimental data set into an intelligent contract data lake, and if the rule is not returned, the transaction is terminated;
and step S30, the experimental data are respectively stored in the MySQL database, the Cache database and the Smart Contract database according to factors such as attributes, characteristics, categories, processing stages and the like. The MySQL database stores all data types of experimental data before the experimental data are subjected to supervision operation, and the Cache database stores data types of short-interval high-frequency fine-grained calling, such as: relationship data, account data, tax data, historical data, scoring data, blacklist \ white list data, and the like. The Smart Contract database stores the data type of the transaction characteristic attribute;
step S40, under the command of the intelligent controller, the characteristic data of different areas on the intelligent contract data are sequentially transmitted to the machine learning engine to execute the operations of early supervision, middle supervision and later supervision, which respectively correspond to the following steps: KYC (customer identification), AML (anti-money laundering test), Credit grading. The intelligent controller firstly calls account data, score data, blacklist/white list and other data cached in a Cache database, and performs KYC operation on the data. If the execution result passes, step S50 is entered, otherwise, the transaction is terminated;
and step S50, under the direction of the intelligent controller, the characteristic attribute data in the Smart Contract database is executed with AML operation, and the transaction data is judged after arbitration judgment of triple methods of behavior modeling, link analysis and anomaly detection to determine the execution result. If the execution result passes, step S60 is entered, otherwise, the transaction is terminated;
and step S60, under the direction of the intelligent controller, the data passed in the step S50 and the characteristic attribute data mapped in the Smart Contract database are subjected to Credit grading operation, and the transaction data obtains corresponding scores in the evaluation of the grading card model to be used as the Credit score of the transaction. The intelligent controller stores the grading result in a Cache database;
and step S70, the intelligent controller returns the final judgment result of the machine learning engine to the Smart Contract database to determine the final transaction result and display the transaction condition and the prediction accuracy.
2. The model of claim 1, wherein the data preprocessing method in step S10 is as follows: and checking data by adopting a head () method, processing missing data, correspondingly adding a default value, deleting incomplete rows and columns, normalizing data types and storing results.
3. The model of claim 1, wherein in step S20, the Oraclize president machine calls different layers from bottom to top in sequence to perform the checking operation, and the logic structure is:
the network protocol is a network topology structure of a centralized prediction machine, and a single centralized service provider controls an intermediate node;
the execution of smart contracts and data calls in the operational layer are both performed on Trusted Execution Environments (TEEs). The AWS takes the audit role and integrity is verified by tlsnottary Proof. Relying on a multi-signature mechanism to allow speakers (Oracles) meeting more than the minimum number of honest nodes to simultaneously sign corresponding nodes;
the contract layer includes order matching contracts, service request contracts, data invocation interfaces, and service standard protocols.
4. The model as claimed in claim 1, wherein in step S40, "the intelligent controller first retrieves the account data, score data and blacklist/whitelist data cached in the Cache database, and performs KYC operation on them", the method is:
step S401, performing Digital organizing and SWIFT filtering operation on the input data, if the result is positive, entering step S402, and if the result is negative, returning, namely terminating the transaction;
step S402, carrying out CDD (customer due diligence) and EDD (enhanced due diligence) operation on the incoming data, if the result is positive, entering step S403, and if the result is negative, returning, namely terminating the transaction;
in step S403, Whitelist/Blacklist Filter operation is performed on the incoming data, and if the result passes, step S50 is performed, otherwise, the transaction is returned, i.e., terminated.
5. A financial transaction supervision model based on intelligent contract data lake according to claim 1, characterized in that step S50 "transaction data determines execution result after arbitration decision of triple method of behavior modeling, link analysis and anomaly detection", its method is:
step S501, an SVM (Support Vector Machine) algorithm is adopted to conduct behavior modeling three-classification operation on transaction data, and the results are as follows: secure transactions, suspicious transactions, pending transactions. If the result is a secure transaction, go to step S503; if the result is suspicious transaction, returning, namely terminating the transaction; if the result is a pending transaction, the step S502 is executed; the SVM algorithm selects a Sigmoid kernel function, and the calculation method comprises the following steps:
wherein X1，X2Is data corresponding to two classes, κ (X)1,X2) Is a sufficient condition for positive definite core, a is used to set gamma parameter setting in kernel function, default is 1/k (k is number of classes), b is used to set coef0 in kernel function, default is 0;
step S502, link analysis operation is carried out on the transaction to be determined by adopting a MaxEnt algorithm (Max entry, maximum Entropy), and if the classification result is safe transaction, the step S503 is carried out; if the result is suspicious transaction, returning, namely terminating the transaction;
step S503, adopt
The Bayesian algorithm detects the abnormality of the secure transaction, if the result is positive, the process goes to step S60, and if the result is negative, the process returns, i.e., the transaction is terminated.
6. A financial transaction supervisory model based on smart contract data lake according to claim 1, wherein step S60 "transaction data gets corresponding score in the evaluation of score card model" by:
step S601, a bucket dividing method is adopted, each processing value is endowed with a corresponding attribute, and numerical value characteristics are converted into classification characteristics;
step S602, calculating an evidence weight (WoE) of each attribute and an Information Value (IV) of each feature point, where the calculation formula of the evidence weight is:
[ln(Distr G/Distr B)]×100
wherein, G represents that the customer transaction passes the target variable being 0, B represents that the customer transaction refutes the target variable being 1;
the information value calculation formula is as follows:
wherein, G represents that the customer transaction passes the target variable being 0, B represents that the customer transaction refutes the target variable being 1;
step S603, model building is performed by replacing the value of the original variable with WoE, and model selection LR (Logistic regression) has the expression:
where y is the probability that the label is A, x is the predicted label, w is the training parameter, w is the probability that the label is ATIs the weight;
step S604, adopting cross validation and grid search to adjust parameters, converting the parameters into two classification problems, wherein the loss function is as follows:
wherein F (w) is a loss function, n is a sample number, ynIs a sample label, p is the corresponding probability;
step S605, calculating a score coefficient of the score card for each attribute to obtain a final score card. The scoring formula is:
where β is an LR coefficient of a given attribute, α is an LR intercept, WoE is an evidence weight of the given attribute, n is a model feature number, Factor, and Offset are scaling parameters.
7. A financial transaction supervisory model based on intelligent contract data lake according to claim 1, wherein the intelligent contract processing method of step S70 is:
step S701, checking whether the information submitted when the user is established and the transaction amount are true and legal;
step S702, checking whether the transfer initiator and the beneficiary are legal users;
and step S703, judging whether the contract operation is continuously executed according to the result returned by the machine learning engine.
8. The intelligent contract data lake-based financial transaction supervision model according to claim 1, wherein the financial transaction supervision model prediction accuracy information can be displayed by using a bar graph.
9. A financial transaction supervisory model based on an intelligent contract data lake, comprising: the system comprises a data deep processing module, a feature marking module, a prediction machine module, an intelligent contract data lake module, a machine learning engine module, an intelligent controller module and an accuracy rate display module;
the data deep processing module is configured to carry out data cleaning data preprocessing operation on the machine learning data set sorted out from the UCI database, and store the result as a transaction data set;
the data feature marking module is configured to construct a six-dimensional feature data set from the transaction data set;
the Oraclize presbyope module is configured to query and check the transaction data through the Oraclize presbyope, judge whether the transaction data are in compliance and then execute the next operation;
the intelligent Contract data lake module is configured to store the transaction data in a MySQL database, a Cache database and a Smart Contract database respectively according to different transaction data attributes, characteristics, types and processing stages;
the machine learning engine module is configured to transmit the feature data of different positions to the machine learning engine in sequence to execute early supervision, middle supervision and later supervision operations;
the intelligent controller module is configured to enable the intelligent controller to have the function of unified command data, algorithm, block and database combined operation;
and the accuracy rate display module is configured to display a final transaction structure, a transaction condition and a prediction accuracy rate.
10. A storage device having a program stored therein, the program being adapted to be loaded and executed by a processor to implement a smart contract data lake based financial transaction supervisory model as claimed in any one of claims 1 to 8.
11. A treatment apparatus comprises
A processor adapted to execute a program, an
A memory adapted to store the program;
wherein the program is adapted to be loaded and executed by a processor to perform:
a financial transaction supervisory model based on smart contract data lakes as claimed in any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110084721.5A CN113011973B (en) | 2021-01-21 | 2021-01-21 | Method and equipment for financial transaction supervision model based on intelligent contract data lake |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110084721.5A CN113011973B (en) | 2021-01-21 | 2021-01-21 | Method and equipment for financial transaction supervision model based on intelligent contract data lake |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113011973A true CN113011973A (en) | 2021-06-22 |
| CN113011973B CN113011973B (en) | 2023-08-29 |
Family
ID=76384624
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110084721.5A Active CN113011973B (en) | 2021-01-21 | 2021-01-21 | Method and equipment for financial transaction supervision model based on intelligent contract data lake |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113011973B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113706154A (en) * | 2021-08-12 | 2021-11-26 | 支付宝(杭州)信息技术有限公司 | Transaction risk detection method, device and equipment |
| CN114565443A (en) * | 2022-04-28 | 2022-05-31 | 深圳高灯计算机科技有限公司 | Data processing method, data processing device, computer equipment and storage medium |
| US11436606B1 (en) * | 2014-10-31 | 2022-09-06 | Experian Information Solutions, Inc. | System and architecture for electronic fraud detection |
| CN115082076A (en) * | 2022-07-04 | 2022-09-20 | 北京天德科技有限公司 | Three-stage financial violation multiple judgment method based on block chain |
| CN115170139A (en) * | 2022-07-04 | 2022-10-11 | 北京天德科技有限公司 | Three-stage financial violation multi-referee system based on block chain data lake |
| US11580259B1 (en) | 2017-09-28 | 2023-02-14 | Csidentity Corporation | Identity security architecture systems and methods |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107085812A (en) * | 2016-12-06 | 2017-08-22 | 雷盈企业管理（上海）有限公司 | The anti money washing system and method for block chain digital asset |
| CN108537667A (en) * | 2018-04-09 | 2018-09-14 | 深圳前海微众银行股份有限公司 | Financial asset anti money washing management-control method, equipment and storage medium based on block chain |
| US20190236598A1 (en) * | 2018-01-31 | 2019-08-01 | Salesforce.Com, Inc. | Systems, methods, and apparatuses for implementing machine learning models for smart contracts using distributed ledger technologies in a cloud based computing environment |
| CN110210968A (en) * | 2019-05-21 | 2019-09-06 | 北京航空航天大学 | Intelligent Service transaction system |
| WO2020102395A1 (en) * | 2018-11-14 | 2020-05-22 | C3.Ai, Inc. | Systems and methods for anti-money laundering analysis |
| US20200167860A1 (en) * | 2018-11-22 | 2020-05-28 | Maria E. Lau | Automated Anti-Money Laundering Compliance SaaS |
| CN111667368A (en) * | 2020-05-29 | 2020-09-15 | 中国工商银行股份有限公司 | Anti-money laundering monitoring system and method |
| CN111768305A (en) * | 2020-06-24 | 2020-10-13 | 中国工商银行股份有限公司 | Anti-money laundering identification method and device |
-
2021
- 2021-01-21 CN CN202110084721.5A patent/CN113011973B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107085812A (en) * | 2016-12-06 | 2017-08-22 | 雷盈企业管理（上海）有限公司 | The anti money washing system and method for block chain digital asset |
| US20190236598A1 (en) * | 2018-01-31 | 2019-08-01 | Salesforce.Com, Inc. | Systems, methods, and apparatuses for implementing machine learning models for smart contracts using distributed ledger technologies in a cloud based computing environment |
| CN108537667A (en) * | 2018-04-09 | 2018-09-14 | 深圳前海微众银行股份有限公司 | Financial asset anti money washing management-control method, equipment and storage medium based on block chain |
| WO2020102395A1 (en) * | 2018-11-14 | 2020-05-22 | C3.Ai, Inc. | Systems and methods for anti-money laundering analysis |
| US20200167860A1 (en) * | 2018-11-22 | 2020-05-28 | Maria E. Lau | Automated Anti-Money Laundering Compliance SaaS |
| CN110210968A (en) * | 2019-05-21 | 2019-09-06 | 北京航空航天大学 | Intelligent Service transaction system |
| CN111667368A (en) * | 2020-05-29 | 2020-09-15 | 中国工商银行股份有限公司 | Anti-money laundering monitoring system and method |
| CN111768305A (en) * | 2020-06-24 | 2020-10-13 | 中国工商银行股份有限公司 | Anti-money laundering identification method and device |
Non-Patent Citations (4)
| Title |
|---|
| NIKOLAOS KAPSOULIS ET AL.: "Know Your Customer (KYC) Implementation with Smart Contracts on a Privacy-Oriented Decentralized Architecture", FUTURE INTERNET, no. 2, pages 1 - 13 * |
| 焦经川;: "区块链与法律的互动:挑战、规制与融合", 云南大学学报(社会科学版), no. 03 * |
| 蔡维德: "互联链：一种新的系统结构和应用构建方法", pages 1 - 16, Retrieved from the Internet <URL:https://www.cnfin.com/upload-xh08/2020/0811/159713240065.pdf> * |
| 郭艳 等: "STO:重新定义证券与泛金融工具的发轫", 经济研究参考, no. 17, pages 59 - 72 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11436606B1 (en) * | 2014-10-31 | 2022-09-06 | Experian Information Solutions, Inc. | System and architecture for electronic fraud detection |
| US11941635B1 (en) | 2014-10-31 | 2024-03-26 | Experian Information Solutions, Inc. | System and architecture for electronic fraud detection |
| US11580259B1 (en) | 2017-09-28 | 2023-02-14 | Csidentity Corporation | Identity security architecture systems and methods |
| CN113706154A (en) * | 2021-08-12 | 2021-11-26 | 支付宝(杭州)信息技术有限公司 | Transaction risk detection method, device and equipment |
| CN114565443A (en) * | 2022-04-28 | 2022-05-31 | 深圳高灯计算机科技有限公司 | Data processing method, data processing device, computer equipment and storage medium |
| CN114565443B (en) * | 2022-04-28 | 2022-09-27 | 深圳高灯计算机科技有限公司 | Data processing method, data processing device, computer equipment and storage medium |
| CN115082076A (en) * | 2022-07-04 | 2022-09-20 | 北京天德科技有限公司 | Three-stage financial violation multiple judgment method based on block chain |
| CN115170139A (en) * | 2022-07-04 | 2022-10-11 | 北京天德科技有限公司 | Three-stage financial violation multi-referee system based on block chain data lake |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113011973B (en) | 2023-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113011973A (en) | Financial transaction supervision model, system and equipment based on intelligent contract data lake | |
| Calderon et al. | A roadmap for future neural networks research in auditing and risk assessment | |
| CN112053221A (en) | Knowledge graph-based internet financial group fraud detection method | |
| WO2010037030A1 (en) | Evaluating loan access using online business transaction data | |
| CN105308640A (en) | Methods and systems for automatically generating high quality adverse action notifications | |
| CN107679997A (en) | Method, apparatus, terminal device and storage medium are refused to pay in medical treatment Claims Resolution | |
| CN106779278A (en) | The evaluation system of assets information and its treating method and apparatus of information | |
| CN112767136A (en) | Credit anti-fraud identification method, credit anti-fraud identification device, credit anti-fraud identification equipment and credit anti-fraud identification medium based on big data | |
| CN112053222A (en) | Knowledge graph-based internet financial group fraud detection method | |
| Plaksiy et al. | Applying big data technologies to detect cases of money laundering and counter financing of terrorism | |
| KR102113347B1 (en) | Method, apparatus and computer program for classifying cryptocurrency accounts using artificial intelligence | |
| Akinbowale et al. | The integration of forensic accounting and big data technology frameworks for internal fraud mitigation in the banking industry | |
| Li et al. | Theory and application of artificial intelligence in financial industry | |
| Barman et al. | A complete literature review on financial fraud detection applying data mining techniques | |
| CN112200583B (en) | Knowledge graph-based fraudulent client identification method | |
| CN112801780A (en) | Method, device and system for identifying international and international risk customers based on federal learning | |
| Ka et al. | Performance Analysis of KN earest Neighbor Classification Algorithms for Bank Loan Sectors | |
| CN117132383A (en) | Credit data processing method, device, equipment and readable storage medium | |
| CN116739764A (en) | Transaction risk detection method, device, equipment and medium based on machine learning | |
| WO2022143431A1 (en) | Method and apparatus for training anti-money laundering model | |
| El-Bannany et al. | Prediction of financial statement fraud using machine learning techniques in UAE | |
| CN109636572A (en) | Risk checking method, device, equipment and the readable storage medium storing program for executing of bank card | |
| CN115564591A (en) | Financing product determination method and related equipment | |
| CN115496364A (en) | Method and device for identifying heterogeneous enterprises, storage medium and electronic equipment | |
| CN114708090A (en) | Bank payment business risk identification device based on big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |