CN116596539A - Money backwashing method and system - Google Patents
Money backwashing method and system Download PDFInfo
- Publication number
- CN116596539A CN116596539A CN202310416603.9A CN202310416603A CN116596539A CN 116596539 A CN116596539 A CN 116596539A CN 202310416603 A CN202310416603 A CN 202310416603A CN 116596539 A CN116596539 A CN 116596539A
- Authority
- CN
- China
- Prior art keywords
- account
- money laundering
- transaction
- money
- wind control
- 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
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000011001 backwashing Methods 0.000 title claims description 6
- 238000004900 laundering Methods 0.000 claims abstract description 210
- 238000012546 transfer Methods 0.000 claims description 21
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 238000012502 risk assessment Methods 0.000 claims description 15
- 230000002776 aggregation Effects 0.000 claims description 14
- 238000004220 aggregation Methods 0.000 claims description 14
- 238000005406 washing Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 4
- 150000001875 compounds Chemical class 0.000 claims description 3
- 238000009792 diffusion process Methods 0.000 claims description 3
- 238000011010 flushing procedure Methods 0.000 claims description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 3
- 238000001228 spectrum Methods 0.000 claims 1
- 230000008859 change Effects 0.000 abstract description 6
- 238000004590 computer program Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- 230000001502 supplementing effect Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 101100421781 Arabidopsis thaliana SNL5 gene Proteins 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q20/401—Transaction verification
- G06Q20/4016—Transaction verification involving fraud or risk level assessment in transaction processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
- G06N5/025—Extracting rules from data
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Accounting & Taxation (AREA)
- Computational Linguistics (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Finance (AREA)
- Computer Security & Cryptography (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Databases & Information Systems (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The application relates to the technical field of money laundering, in particular to a money laundering method and a money laundering system. According to the money laundering method, the wind control knowledge graph is constructed according to historical transaction data, depth-first search is conducted based on the wind control knowledge graph to automatically find out wind control rules, a wind control rule base is formed by the wind control knowledge graph and the wind control rules constructed manually in the prior art, money laundering possibility of an account is calculated according to the wind control rule base, money laundering risks are estimated, change of money laundering means can be adapted in time while rule management and application efficiency are improved, and money laundering risk estimation accuracy is improved.
Description
Technical Field
The application relates to the technical field of money laundering, in particular to a money laundering method and a money laundering system.
Background
With the acceleration of the development process of the Internet and the massive application of financial science and technology, money laundering means are more hidden and complex, money laundering behaviors are characterized by high organization degree, large number of participants and dense funds transfer network, and the requirement on the effectiveness of money laundering monitoring is increasing.
However, in the face of gradual grouping, large-scale and specialized money laundering, the crime working methods of money laundering are continuously changed, the traditional mode of manually constructing wind control rules and blacklists is based, the setting of the used money laundering rules and abnormal transaction index thresholds is relatively fixed, the setting is excessively dependent on historical data and expert experience, novel money laundering characteristics cannot be captured, and rapid changes of money laundering methods and modes are difficult to timely deal with.
Disclosure of Invention
Therefore, the application aims to solve the technical problems that the rule updating is not timely, the quick change of the money laundering method and the mode is difficult to deal with in time, and the money laundering risk assessment accuracy is low in the prior art.
In order to solve the technical problems, the application provides a money laundering method, which comprises the following steps:
calculating the money laundering possibility of the account according to an air control rule base, wherein the air control rule base comprises an air control rule constructed manually and an air control rule automatically found by depth-first search based on an air control knowledge graph, and the air control knowledge graph is constructed based on historical transaction data;
and if the money laundering possibility is larger than a first preset threshold value, the account is a money laundering account, otherwise, the account is a normal account.
Preferably, the wind control knowledge graph is constructed based on historical transaction data, and specifically comprises:
and associating transaction running water, account information and client information in the historical transaction data in a form of a knowledge graph, and processing the historical transaction data according to transaction-level indexes, account-level indexes and client-level indexes to form an account general portrait.
Preferably, the wind control rule automatically discovered by the depth-first search based on the wind control knowledge graph comprises a compound rule describing transitive logic in the wind control knowledge graph:
isfraud(X,Y)←b 1 (X,A 1 )∧b 2 (A 1 ,A 2 )∧…∧b n (A n-1 ,Y)
and attribute rules describing relationship or entity attribute presence logic in the wind-controlled knowledge graph:
isfraud(X,fraud)←b 1 (X,a ′ )
wherein a is ′ Representing a specific entity, A i Representing variables, b i Representing the relationship, isfraud represents whether the account is laundered, X represents the account to be evaluated, Y ε { normal, fraud } represents the laundering label, normal represents the normal account, fraud represents the laundering account.
Preferably, the calculation formula for calculating the money laundering possibility of the account according to the wind control rule base is as follows:
where Prob (X) represents the likelihood that account X is a money laundering account, r i Representing the money laundering association rules satisfied by account X, conf (r i ) Is the confidence of the rule.
Preferably, the money back-washing method further comprises:
and identifying the money-washing account and the money-washing partner according to the node characteristics and the path characteristics of the spectrogram structure of the wind control knowledge graph.
Preferably, the identifying the money laundering account and the money laundering partner according to the node characteristics and the path characteristics of the spectrogram structure of the wind control knowledge graph comprises:
calculating the transaction aggregation degree of the account according to the account node transfer transaction relation and the transaction weights of the transaction parties:
wherein, agg (X) i ) Representing the transaction aggregation level of the account, d ij E {0,1} indicates whether there is a transfer transaction between account i and account j, w ij For transferring accounts for both partiesTransaction weight, lambda E [0,1 ]]Is a balance factor;
calculating the comprehensive money laundering risk degree of the account according to the money laundering possibility of the account and the transaction aggregation degree:
Risk(X i )＝βAgg(X i )+(1-β)Prob(X i ),i＝1～n
wherein, risk (X) i ) Representing the comprehensive money laundering risk of the account, agg (X i ) Representing the transaction aggregation level of the account, prob (X i ) Representing the money laundering possibility of the account, βε [0,1 ]]Is a balance factor;
if the comprehensive money laundering risk of the account is larger than a second preset threshold, the account is a money laundering account, otherwise, the account is a normal account.
Preferably, the identifying the money laundering account and the money laundering partner further comprises:
according to the association relation among account nodes, carrying out money laundering risk propagation and diffusion, and calculating the money laundering risk degree after the account propagation:
wherein Prob (X i ) For account X calculated from a wind control rule base i The money laundering possibility, w ij For both parties to transfer transaction weights, trans (X i ) Representing account X i The risk of money laundering after transmission;
and if the risk degree of the back money laundering after the propagation is larger than a third preset threshold value, the account is a money laundering account, otherwise, the account is a normal account.
Preferably, the identifying the money laundering account and the money laundering partner further comprises:
the discovered money laundering accounts are related by using a community discovery algorithm, and transaction groups are divided, wherein the calculation formula of the overall money laundering risk degree of the group partner of the transaction groups is as follows:
wherein, groupRisk (C) i ) Representing transaction group C i Overall money laundering risk of (2), X ij For transaction group C i Is a member account of Risk (X ij ) Representing member account X ij Is a comprehensive money laundering risk degree, w ij For member account X ij In transaction group C i Importance weights of (2);
and if the overall money laundering risk of the group is greater than a fourth preset threshold, the transaction group is a money laundering group.
The application also provides a money laundering system comprising:
the wind control knowledge graph management module is used for constructing and managing a wind control knowledge graph;
the knowledge reasoning model management module is used for constructing and managing a wind control rule base, wherein the wind control rule base comprises wind control rules constructed manually and wind control rules automatically found by depth-first search based on a wind control knowledge graph;
and the money laundering risk assessment module is used for calculating the money laundering possibility of the account according to the wind control rule base, if the money laundering possibility is larger than a first preset threshold value, the account is a money laundering account, and otherwise, the account is a normal account.
Preferably, the money laundering risk assessment module is further used for identifying money laundering accounts and money laundering partners according to node characteristics and path characteristics of the spectrogram structure of the wind control knowledge graph.
Compared with the prior art, the technical scheme of the application has the following advantages:
according to the money laundering method, the wind control knowledge graph is constructed according to historical transaction data, depth-first search is conducted based on the wind control knowledge graph to automatically find out wind control rules, a wind control rule base is formed by the wind control knowledge graph and the wind control rules constructed manually in the prior art, money laundering possibility of an account is calculated according to the wind control rule base, money laundering risks are estimated, change of money laundering means can be adapted in time while rule management and application efficiency are improved, and money laundering risk estimation accuracy is improved.
Drawings
In order that the application may be more readily understood, a more particular description of the application will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings, in which:
FIG. 1 is a flow chart of an embodiment of a money laundering method according to the present application;
FIG. 2 is a schematic diagram of a wind-controlled knowledge graph according to an embodiment of the present application;
FIG. 3 is a general flow chart of a money laundering method according to an embodiment of the present application;
FIG. 4 is a schematic diagram of a money laundering system according to the present application.
Detailed Description
The core of the application is to provide a money laundering method and a money laundering system, which can adapt to the change of money laundering means in time while improving the rule management and application efficiency, and improve the accuracy of money laundering risk assessment.
In order to better understand the aspects of the present application, the present application will be described in further detail with reference to the accompanying drawings and detailed description. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
Referring to fig. 1, fig. 1 is a flowchart of an implementation of a money back flushing method according to the present application; the specific operation steps are as follows:
s101, calculating money laundering possibility of an account according to an air control rule base, wherein the air control rule base comprises an air control rule constructed manually and an air control rule automatically found by depth-first search based on an air control knowledge graph, and the air control knowledge graph is constructed based on historical transaction data;
s102, if the money laundering possibility is larger than a first preset threshold value, the account is a money laundering account, otherwise, the account is a normal account.
According to the money laundering method, the wind control knowledge graph is constructed according to historical transaction data, depth-first search is conducted based on the wind control knowledge graph to automatically find out wind control rules, a wind control rule base is formed by the wind control rules constructed manually in the prior art, money laundering possibility of an account is calculated according to the wind control rule base, money laundering risks are estimated, change of money laundering means can be adapted in time while rule management and application efficiency are improved, and money laundering risk estimation accuracy is improved.
Based on the above embodiment, the embodiment constructs a hierarchical wind control knowledge graph based on commercial bank client data and hierarchical wind control indexes to form a general representation of a bank account, specifically:
and correlating transaction running water, account information and client information in the historical transaction data in the form of a knowledge graph, and processing the historical transaction data according to the transaction-level index, the account-level index and the client-level index to form an account general portrait.
The wind control knowledge graph consists of three layers of graphs from bottom to top: the first layer is an example layer, and a client data layer directly constructed by the original data organizes transaction flow, account information and client information in a form of a knowledge graph; the second layer is a conceptual layer, different levels of different types of wind control indexes are set for processing the original data, and a general portrait layer formed by the wind control indexes is formed; and the third layer is an reasoning layer, and based on account portrait information, rule reasoning and graph structure reasoning are fused to carry out comprehensive assessment of money laundering risks.
As in fig. 2, the instance-layer knowledge graph is the account information, customer information, and transaction flow represented in the form of triples or tuples, such as: (account, open account, bank, date of open account), (account, holder, customer), (customer, age value), (customer, loan balance, amount value), (account, whether social security card is bound, yes/no), (customer, whether to purchase financial product, yes/no), (account, transfer, account, transaction time, transaction amount, transaction facility, lending direction, current transfer identifier, pre-transaction account balance, post-transaction account balance), etc. The concept layer knowledge graph is an account portrait expressed in a form of a triplet or a multi-element group, such as: (account, large transfer, account), (account, average daily account balance, value of amount), (account, suspicious transaction number, corresponding index rating), (account, average daily transaction amount, corresponding index rating), (account, number of transaction partners, corresponding index rating), (account, average transaction amount, corresponding index rating), (customer, age group) and the like. The inference layer knowledge graph is a risk assessment process expressed in a multi-element form, gives the possibility of money laundering of an account, identifies a money laundering account, and discovers money laundering partners, such as: (account, risk assessment, money laundering/normal, wind control index, wind control value, confidence), (transaction group, risk assessment, money laundering/normal, wind control index, wind control value, confidence) etc.
Based on the above embodiments, the pneumatic control knowledge graph stores account portrait data and inter-account transaction data which are integrated by raw data extraction, wherein potential money laundering features and money laundering modes are included. The money laundering evaluation and decision made based on the wind control rule has stronger interpretation, but the wind control rule needs to be updated in time according to the change of the money laundering mode. Therefore, the application carries out thorough depth-first search on the constructed wind control knowledge graph by introducing an efficient and determined rule mining and filtering algorithm, thereby realizing the automatic discovery of the wind control rule.
The wind control rule used in the application follows the format of the traditional wind control rule, consists of wind control indexes, threshold values and confidence, and is a subset of the Horn rule form, and specifically utilizes the following 2 general forms as follows:
isfraud(X,Y)←b 1 (X,A 1 )∧b 2 (A 1 ,A 2 )∧…∧b n (A n-1 ,Y)
isfraud(X,fraud)←b 1 (X,a ′ )#(1)
wherein a is ′ Representing a specific entity, A i Representing variables, b i Representing the relationship, isfraud represents whether the account is laundered, X represents the account to be evaluated, Y ε { normal, fraud } represents the laundering label, normal represents the normal account, fraud represents the laundering account. Rule 1 is called a compound rule, which describes transitive logic in the knowledge graph, such as that the account with the money-laundering account having a transfer transaction chain is a money-laundering account. The 2 nd rule is called attribute rule, which describes the existence logic of relation or entity attribute in the wind control knowledge graph, for example, A has a large number of suspicious money laundering transactions, then it is a money laundering account. The confidence of each rule describes the probability that this rule holds, and for one rule r, the confidence is defined as the probability that it holds on the training knowledge graph.
In one embodiment, the automatically discovered wind control rules are as follows:
isfraud(X,Y)←TAMTL3(X,A 1 )∧TAMTL2(A 1 ,A 2 )∧isfraud(A 2 ,Y)
isfraud(X,fraud)←suspicious_num(X,SNL5)
the meaning of the first rule is if account X is towards account A 1 Transferring accounts with three-level transfer amount (714-9789 yuan), and continuing to transfer A1 to account A 2 Transferring money and the transfer amount is two-level (0.9-714 yuan), then X and A2 are both money-laundering accounts or normal accounts, and the confidence of the rule is 0.7;
the second rule has the meaning that if the number of suspicious transactions for account X is five (more than 50), then X is a money laundering account with a confidence of 0.82.
Based on the above embodiment, the calculation formula for calculating the money laundering possibility of the account according to the wind control rule base is as follows:
where Prob (X) represents the likelihood that account X is a money laundering account, r i Representing the money laundering association rules satisfied by account X, conf (r i ) Is the confidence of the rule.
And determining an optimal classification threshold value by taking the maximized F1 score as an unbiased measure, and considering the account with the money laundering possibility larger than the threshold value as a money laundering account, otherwise, determining the account as a normal account.
The money backwashing scene requires a model with strong interpretability, most of various neural network models used based on the reasoning method representing learning have complex operation processes, have strong black box attributes naturally, and are difficult to form persuasive analysis reports for decision reference of a supervision organization according to the operation processes. The reasoning model based on the graph structure can fully utilize graph attributes of the wind control knowledge graph, and node characteristics and path characteristics in the graph structure are mined, so that money laundering accounts and money laundering partners are identified. Based on the constructed wind control knowledge graph, the application uses two social network analysis methods (SNA) and a risk propagation model for discovering network node importance and communities to identify potential money laundering nodes and money laundering partners, thereby improving the identification accuracy and ensuring better interpretability of the model.
Network node importance measure: calculating transaction aggregation degree of the account by using a network node importance measurement algorithm, and supplementing a money laundering account identification result according to the determined optimal classification threshold, wherein the method comprises the following steps of:
the application refers to the feature vector centrality and PageRank algorithm thought, and evaluates the transaction aggregation degree of the accounts according to the transfer transaction relationship among the accounts and the importance of transaction opponents, and specifically adopts the following formula for calculation, wherein the formula (4) is an iteration initial value.
Wherein, agg (X) i ) Representing the transaction aggregation level of the account, d ij E {0,1} indicates whether there is a transfer transaction between account i and account j, w ij For the weight of the transfer transaction of the two parties, the transfer times and the transfer money between the two parties are usedQuota determination, lambda E [0,1 ]]To balance the factors, the account itself and the transaction opponents balance the impact on the transaction aggregation level measure.
Further, the comprehensive money laundering risk degree of the account is calculated according to the money laundering possibility of the account and the transaction aggregation degree:
Risk(X i )＝βAgg(X i )+(1-β)Prob(X i ),i＝1～n#(5)
wherein, risk (X) i ) Representing the comprehensive money laundering risk of the account, agg (X i ) Representing the transaction aggregation level of the account, prob (X i ) Representing the money laundering possibility of the account, βε [0,1 ]]And as a balance factor, balancing the influence of the rule reasoning result and the graph structure reasoning result on the comprehensive assessment of the money laundering risk.
And finally, determining an optimal classification threshold value by taking the maximized F1 score as an unbiased measure, and considering the account with the comprehensive money laundering risk degree larger than the threshold value as a money laundering account, or else, determining the account as a normal account.
Risk propagation: and (3) using a risk propagation algorithm to obtain the post-propagation money laundering risk degree of the account, and supplementing a money laundering account identification result according to the determined optimal classification threshold, wherein the recognition result is as follows:
according to the application, the propagation and diffusion of the money laundering risks are carried out according to the association relation among the accounts, the evaluation of the money laundering risks of the auxiliary accounts is carried out, and the money laundering risk degree after the propagation of the accounts is calculated by adopting the following formula (6) and formula (7).
Wherein Prob (X i ) For account X calculated from a wind control rule base i The money laundering possibility, w ij For both parties to transfer transaction weights, trans (X i ) Representing account X i Is transmitted from the high probability account to the transaction networkAnd the risk of laundering after transmission is defined as the number of standard deviations of the likelihood of laundering from the average.
And finally, determining an optimal classification threshold value by taking the maximized F1 score as an unbiased measure, and considering the account with the transmitted money laundering risk degree larger than the threshold value as a money laundering account, or else, determining the account as a normal account.
Community discovery: dividing all accounts into different transaction groups by using a community discovery algorithm, calculating the overall money laundering risk of each transaction group, and determining money laundering and grouping according to the determined optimal threshold value, wherein the method comprises the following steps of:
the money laundering behavior is generally represented as a common proposal of a large number of accounts, the internal transaction of a money laundering partner is dense, and the transactions of the money laundering accounts and the accounts outside the partner are rare, so the application is based on the discovered high suspicious money laundering accounts, realizes the excavation of the associated money laundering accounts and money laundering partners by using a community discovery algorithm, and particularly uses a Luwen algorithm, a label propagation algorithm and a modularity algorithm. In the divided communities, the money laundering partners are identified according to the partner overall money laundering risk indexes, and specifically, the following formula (8) is adopted to calculate the partner overall money laundering risk.
Wherein, groupRisk (C) i ) Representing transaction group C i Overall money laundering risk of (2), X ij For transaction group C i Is a member account of Risk (X ij ) Representing member account X ij Is a comprehensive money laundering risk degree, w ij For member account X ij In transaction group C i Importance weight of (c).
And finally, determining an optimal classification threshold value by taking the maximized F1 score as an unbiased measure, and identifying a transaction group with the overall money laundering risk larger than the threshold value as a money laundering partner.
For different transaction groups, supplementing a money laundering partner discovery result according to the number of the identified money laundering accounts in the transaction groups; and according to the discovered money laundering bulk, identifying the intra-bulk account as an associated money laundering account, and supplementing the money laundering account identification result, wherein the method comprises the following steps of:
according to the coverage condition of the high-doubt money laundering account, determining a transaction group with three or more money laundering accounts as a money laundering bulk; the money laundering party internal account is identified as an associated money laundering account.
FIG. 3 is a general flow chart of a money laundering method according to an embodiment of the present application.
As shown in fig. 4, the present application also provides a money laundering system comprising:
the wind control knowledge graph management module is used for constructing and managing a wind control knowledge graph; the module supports knowledge graph ontology management and instance management, and the ontology management is responsible for manual construction of the knowledge graph ontology and comprises adding, deleting and modifying of entity types and relation types; instance management is responsible for manual importation of wind control data, including addition, deletion, and modification of entity instances and relationship instances.
The knowledge reasoning model management module is divided into a rule module and a graph structure module. The rule module is responsible for managing the wind control rules, and comprises manual rule construction, rule inquiry and rule automatic discovery, and is used for constructing and managing a wind control rule base, wherein the wind control rule base comprises manually constructed wind control rules and wind control rules which are automatically discovered by depth-first search based on a wind control knowledge graph; the graph structure module is responsible for managing a knowledge reasoning model based on the graph structure, and comprises an algorithm selection and parameter setting of a node importance measurement algorithm, a community discovery algorithm and a risk propagation algorithm.
And the money laundering risk assessment module provides three functions of money laundering risk assessment, money laundering account identification and money laundering bulk discovery. The money laundering risk assessment is based on the meeting condition of the wind control rule, the node importance is measured, the risk transmission result carries out comprehensive money laundering risk assessment on the account, and the account money laundering possibility or money laundering risk degree is given; the money laundering account identification is based on risk assessment, a certain threshold value is set, an account with the possibility of money laundering or the risk of money laundering being greater than the threshold value is identified as a money laundering account, and meanwhile, the internal account of the money laundering partner can be identified as a money laundering account on the discovery result of the money laundering partner; based on community discovery and risk assessment, the money-washing group discovery sets a certain threshold, considers the transaction group with the overall money-washing risk degree larger than the threshold as the money-washing group, and can simultaneously consider the transaction group where a large number of money-washing accounts are located as the money-washing group on the money-washing group identification result.
The specific embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium is stored with a computer program, and the computer program realizes the steps of the money back-flushing method when being executed by a processor.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It is apparent that the above examples are given by way of illustration only and are not limiting of the embodiments. Other variations and modifications of the present application will be apparent to those of ordinary skill in the art in light of the foregoing description. It is not necessary here nor is it exhaustive of all embodiments. While still being apparent from variations or modifications that may be made by those skilled in the art are within the scope of the application.
Claims (10)
1. A method of backwashing money, comprising:
calculating the money laundering possibility of the account according to an air control rule base, wherein the air control rule base comprises an air control rule constructed manually and an air control rule automatically found by depth-first search based on an air control knowledge graph, and the air control knowledge graph is constructed based on historical transaction data;
and if the money laundering possibility is larger than a first preset threshold value, the account is a money laundering account, otherwise, the account is a normal account.
2. The money back-flushing method according to claim 1, wherein the pneumatic control knowledge graph is constructed based on historical transaction data, and specifically comprises:
and associating transaction running water, account information and client information in the historical transaction data in a form of a knowledge graph, and processing the historical transaction data according to transaction-level indexes, account-level indexes and client-level indexes to form an account general portrait.
3. The method of claim 1, wherein the wind control rules automatically discovered based on the depth-first search of the wind control knowledge graph comprise compound rules describing transitive logic in the wind control knowledge graph:
isfraud(X,Y)←b 1 (X,A 1 )∧b 2 (A 1 ,A 2 )∧…∧b n (A n-1 ,Y)
and attribute rules describing relationship or entity attribute presence logic in the wind-controlled knowledge graph:
isfraud(X,fraud)←b 1 (X,a ′ )
wherein a is ′ Representing a specific entity, A i Representing variables, b i Representing the relationship, isfraud represents whether the account is laundered, X represents the account to be evaluated, Y ε { normal, fraud } represents the laundering label, normal represents the normal account, fraud represents the laundering account.
4. The method of claim 1, wherein the calculation formula for calculating the money laundering probability of the account according to the wind control rule base is:
where Prob (X) represents the likelihood that account X is a money laundering account, r i Representing the money laundering association rules satisfied by account X, conf (r i ) Is the confidence of the rule.
5. The money back-washing method of claim 1, further comprising:
and identifying the money-washing account and the money-washing partner according to the node characteristics and the path characteristics of the spectrogram structure of the wind control knowledge graph.
6. The method of claim 1, wherein the identifying money laundering accounts and money laundering partners based on node features and path features of the wind-controlled knowledge graph spectrogram structure comprises:
calculating the transaction aggregation degree of the account according to the account node transfer transaction relation and the transaction weights of the transaction parties:
wherein, agg (X) i ) Representing the transaction aggregation level of the account, d ij E {0,1} indicates whether there is a transfer transaction between account i and account j, w ij For the transfer transaction weight of both parties, lambda E [0,1]Is a balance factor;
calculating the comprehensive money laundering risk degree of the account according to the money laundering possibility of the account and the transaction aggregation degree:
Risk(X i )＝βAgg(X i )+(1-β)Prob(X i ),i＝1～n
wherein, risk (X) i ) Representing the comprehensive money laundering risk of the account, agg (X i ) Representing the transaction aggregation level of the account, prob (X i ) Representing the money laundering possibility of the account, βε [0,1 ]]Is a balance factor;
if the comprehensive money laundering risk of the account is larger than a second preset threshold, the account is a money laundering account, otherwise, the account is a normal account.
7. The method of claim 1, wherein the identifying money laundering accounts and money laundering partners based on node features and path features of the wind-controlled knowledge graph spectrogram structure further comprises:
according to the association relation among account nodes, carrying out money laundering risk propagation and diffusion, and calculating the money laundering risk degree after the account propagation:
wherein Prob (X i ) For account X calculated from a wind control rule base i The money laundering possibility, w ij For both parties to transfer transaction weights, trans (X i ) Representing account X i The risk of money laundering after transmission; and if the risk degree of the back money laundering after the propagation is larger than a third preset threshold value, the account is a money laundering account, otherwise, the account is a normal account.
8. The method of claim 1, wherein the identifying money laundering accounts and money laundering partners based on node features and path features of the wind-controlled knowledge graph spectrogram structure further comprises:
the discovered money laundering accounts are related by using a community discovery algorithm, and transaction groups are divided, wherein the calculation formula of the overall money laundering risk degree of the group partner of the transaction groups is as follows:
wherein, groupRisk (C) i ) Representing transaction group C i Overall money laundering risk of (2), X ij For transaction group C i Is a member account of Risk (X ij ) Representing member account X ij Is a comprehensive money laundering risk degree, w ij For member account X ij In transaction group C i Importance weights of (2);
and if the overall money laundering risk of the group is greater than a fourth preset threshold, the transaction group is a money laundering group.
9. A money laundering system, comprising:
the wind control knowledge graph management module is used for constructing and managing a wind control knowledge graph;
the knowledge reasoning model management module is used for constructing and managing a wind control rule base, wherein the wind control rule base comprises wind control rules constructed manually and wind control rules automatically found by depth-first search based on a wind control knowledge graph;
and the money laundering risk assessment module is used for calculating the money laundering possibility of the account according to the wind control rule base, if the money laundering possibility is larger than a first preset threshold value, the account is a money laundering account, and otherwise, the account is a normal account.
10. The money laundering system of claim 9 wherein the money laundering risk assessment module is further configured to identify money laundering accounts and money laundering partners based on node features and path features of the wind-controlled knowledge graph spectrum structure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310416603.9A CN116596539B (en) | 2023-04-18 | 2023-04-18 | Money backwashing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310416603.9A CN116596539B (en) | 2023-04-18 | 2023-04-18 | Money backwashing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116596539A true CN116596539A (en) | 2023-08-15 |
| CN116596539B CN116596539B (en) | 2024-02-09 |
Family
ID=87606954
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310416603.9A Active CN116596539B (en) | 2023-04-18 | 2023-04-18 | Money backwashing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116596539B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111476662A (en) * | 2020-04-13 | 2020-07-31 | 中国工商银行股份有限公司 | Anti-money laundering identification method and device |
| CN113256121A (en) * | 2021-05-26 | 2021-08-13 | 杭州云象网络技术有限公司 | Artificial intelligent money laundering method and system |
-
2023
- 2023-04-18 CN CN202310416603.9A patent/CN116596539B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111476662A (en) * | 2020-04-13 | 2020-07-31 | 中国工商银行股份有限公司 | Anti-money laundering identification method and device |
| CN113256121A (en) * | 2021-05-26 | 2021-08-13 | 杭州云象网络技术有限公司 | Artificial intelligent money laundering method and system |
Non-Patent Citations (1)
| Title |
|---|
| 袁俊等: ""知识图谱在商业银行风控领域的研究与应用综述"", 《计算机工程与应用》, vol. 19, no. 58, pages 37 - 48 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116596539B (en) | 2024-02-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Majhi et al. | Fuzzy clustering using salp swarm algorithm for automobile insurance fraud detection | |
| CN110751557B (en) | Abnormal fund transaction behavior analysis method and system based on sequence model | |
| CN110414780B (en) | Fraud detection method based on generation of financial transaction data against network | |
| Kim et al. | Predicting repayment of borrows in peer‐to‐peer social lending with deep dense convolutional network | |
| CN109766454A (en) | A kind of investor's classification method, device, equipment and medium | |
| CN113011973B (en) | Method and equipment for financial transaction supervision model based on intelligent contract data lake | |
| Satyanarayana et al. | Survey of classification techniques in data mining | |
| Fan et al. | Improved ML‐based technique for credit card scoring in Internet financial risk control | |
| CN113657896A (en) | Block chain transaction topological graph analysis method and device based on graph neural network | |
| CN110222733A (en) | The high-precision multistage neural-network classification method of one kind and system | |
| CN106097094A (en) | A kind of man-computer cooperation credit evaluation new model towards medium-sized and small enterprises | |
| Zhang et al. | Consumer credit risk assessment: A review from the state-of-the-art classification algorithms, data traits, and learning methods | |
| Li et al. | An early warning model for customer churn prediction in telecommunication sector based on improved bat algorithm to optimize ELM | |
| CN113343123B (en) | Training method and detection method for generating confrontation multiple relation graph network | |
| Owusu et al. | A deep learning approach for loan default prediction using imbalanced dataset | |
| CN116596539B (en) | Money backwashing method and system | |
| Pendharkar et al. | A misclassification cost‐minimizing evolutionary–neural classification approach | |
| CN116541792A (en) | Method for carrying out group partner identification based on graph neural network node classification | |
| CN115242556B (en) | Network anomaly detection method based on incremental self-encoder | |
| Muthukumaran et al. | Feature Selection with Optimal Variational Auto Encoder for Financial Crisis Prediction. | |
| CN116805245A (en) | Fraud detection method and system based on graph neural network and decoupling representation learning | |
| Zhao et al. | Network-based feature extraction method for fraud detection via label propagation | |
| CN112632219B (en) | Method and device for intercepting junk short messages | |
| CN111984842B (en) | Bank customer data processing method and device | |
| CN117094817B (en) | Credit risk control intelligent prediction method and system |
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 |