Transaction risk control method and apparatus

Abstract

The application discloses a transaction risk control method and device, and relates to the financial field, and the method comprises the following steps: screening out customers including transaction data corresponding to no network and network signals as reference customers based on historical transaction data of bank customers; classifying the reference customers to obtain multiple reference customer categories; determining a to-be-controlled risk type for each reference customer category based on the transaction data corresponding to no network signals; determining a corresponding risk control model for each to-be-controlled risk type based on the transaction data; determining a reference customer category of a customer including only transaction data corresponding to network signals, issuing a risk control model corresponding to a to-be-controlled risk type of the customer to a mobile terminal of the customer, and using the risk control model stored in the mobile terminal to control the transaction of the mobile terminal in a no network signal state. The application can control the transaction in a no network signal state.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of finance, in particular to a transaction risk control method and device. BACKGROUND

[0002] This section is intended to provide background or context to the embodiments of the application recited in the claims. The description herein does not constitute admission that the prior publication, square, or subject matter described herein and / or the material contained therein is or was prior art to the claims at issue.

[0003] At present, the transaction risk control of the bank is deployed on the bank server, so the network signal of the mobile terminal of the customer needs to be good. At present, the risk control of the bank server is based on the online transaction of the customer. When the client is in no network signal or weak network signal, the waiting time of the transaction of the customer will be very long, and even the transaction fails. The bank server does not have corresponding risk control, that is, only the online transaction of the customer is controlled, and the risk control of no network signal cannot be controlled, so that the risk that may be encountered in actual situation cannot be found in advance. The actual effect of the existing risk control is limited. SUMMARY

[0004] The embodiments of the present application provide a transaction risk control method to solve the technical problem that the existing risk control cannot control the risk of no network signal, so that the risk that may be encountered in actual situation cannot be found in advance, and the actual effect of the existing risk control is limited. The method comprises:

[0005] obtain the historical transaction data of the bank customer;

[0006] based on the historical transaction data of the bank customer, filter out the customers including the transaction data corresponding to no network signal and the transaction data corresponding to network signal, and determine the filtered customers as reference customers;

[0007] classify the reference customers to obtain a plurality of reference customer categories;

[0008] for each reference customer category, based on the transaction data corresponding to no network signal of the reference customer category, determine the risk type to be controlled of the reference customer category in no network signal;

[0009] for each risk type to be controlled of the reference customer category, based on the transaction data corresponding to no network signal of the reference customer category, determine the risk control model corresponding to the risk type to be controlled;

[0010] for the customer only including the transaction data corresponding to network signal, determine the reference customer category corresponding to the customer;

[0011] The risk control model corresponding to the to-be-controlled risk type of the reference customer category corresponding to the customer is issued to the mobile terminal of the customer, and when the mobile terminal of the customer is in a no-network-signal state, the transaction of the mobile terminal is controlled by using the risk control model stored in the mobile terminal.

[0012] The embodiment of the application further provides a transaction risk control device for solving the technical problem that the existing risk control cannot control risk in a no-network-signal state, actual risks that may be encountered in practice cannot be found in advance, and the actual effect of the existing risk control is limited, and the device comprises:

[0013] A data acquisition module is configured to acquire historical transaction data of a bank customer.

[0014] A data screening module is configured to screen out customers including transaction data corresponding to a no-network-signal state and transaction data corresponding to a network-signal state based on the historical transaction data of the bank customer, and determine the screened-out customers as reference customers.

[0015] A reference customer classification module is configured to classify the reference customers to obtain a plurality of reference customer categories.

[0016] A risk type determination module is configured to determine, for each reference customer category, a to-be-controlled risk type of the reference customer category in a no-network-signal state based on transaction data corresponding to the no-network-signal state of the reference customer category.

[0017] A risk control model determination module is configured to determine, for each to-be-controlled risk type of the reference customer category, a risk control model corresponding to the to-be-controlled risk type based on transaction data corresponding to the no-network-signal state of the reference customer category.

[0018] The reference customer classification module is further configured to determine the reference customer category corresponding to a customer including only transaction data corresponding to a network-signal state.

[0019] A risk control model issuing module is configured to issue the risk control model corresponding to the to-be-controlled risk type of the reference customer category corresponding to the customer to the mobile terminal of the customer, and control the transaction of the mobile terminal by using the risk control model stored in the mobile terminal when the mobile terminal of the customer is in a no-network-signal state.

[0020] The embodiment of the application further provides a computer device, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the transaction risk control method when executing the computer program.

[0021] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the transaction risk control method.

[0022] The embodiment of the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to realize the transaction risk control method.

[0023] Compared with the prior art, in the embodiment of the present application, the historical transaction data of the bank customer is obtained, the customers including the transaction data corresponding to the no-network signal and the transaction data corresponding to the network signal are screened out based on the historical transaction data of the bank customer, the screened-out customers are determined as reference customers, the reference customers are classified to obtain a plurality of reference customer categories, for each reference customer category, the risk type to be controlled of the reference customer category in the no-network signal is determined based on the transaction data corresponding to the no-network signal of the reference customer category, for each risk type to be controlled of the reference customer category, the risk control model corresponding to the risk type to be controlled is determined based on the transaction data corresponding to the no-network signal of the reference customer category, for the customer including only the transaction data corresponding to the network signal, the reference customer category corresponding to the customer is determined, the risk control model corresponding to the risk type to be controlled of the reference customer category corresponding to the customer is issued to the mobile terminal of the customer, and when the mobile terminal of the customer is in the no-network signal state, the transaction of the mobile terminal is controlled by the risk control model stored in the mobile terminal, so that the transaction can be controlled in the no-network signal state. BRIEF DESCRIPTION OF DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort. In the drawings:

[0025] Figure 1 The transaction risk control method flow in the embodiment of the present application Figure 1 ;

[0026] Figure 2 The transaction risk control method flow in the embodiment of the present application Figure 2 ;

[0027] Figure 3 The transaction risk control method flow in the embodiment of the present application Figure 3 ;

[0028] Figure 4 The transaction risk control method flow for the embodiment of the present application Figure 4 ;

[0029] Figure 5 The transaction risk control method flow for the embodiment of the present application Figure 5 ;

[0030] Figure 6 The structure block diagram of the transaction risk control device for the embodiment of the present application. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, further detailed description of the embodiments of the present application will be given below with reference to the drawings. Herein, the illustrative embodiments of the present application and the description thereof are used to explain the present application, but not as a limitation of the present application.

[0032] The acquisition, storage, use, processing and the like of data in the technical scheme of the present application all comply with the relevant provisions of national laws and regulations.

[0033] Based on the problems existing in the prior art, the present application proposes a transaction risk control method, which is applied to a bank server, and the specific flow is as shown in Figure 1 The method comprises the following steps.

[0034] Step 101: Obtain the historical transaction data of bank customers.

[0035] Step 102: Based on the historical transaction data of bank customers, filter out customers including transaction data corresponding to no network signal and transaction data corresponding to network signal, and determine the filtered customers as reference customers.

[0036] Step 103: Classify the reference customers to obtain multiple reference customer categories.

[0037] Step 104: For each reference customer category, based on the transaction data corresponding to no network signal of the reference customer category, determine the risk type to be controlled of the reference customer category under no network signal.

[0038] Step 105: For each risk type to be controlled of the reference customer category, based on the transaction data corresponding to no network signal of the reference customer category, determine the risk control model corresponding to the risk type to be controlled.

[0039] Step 106: For the customers including only transaction data corresponding to network signal, determine the reference customer category corresponding to the customers.

[0040] Step 107: issuing the risk control model corresponding to the risk type to be controlled of the reference customer category corresponding to the customer to the mobile terminal of the customer, and when the mobile terminal of the customer is in a network signal-free state, using the risk control model stored in the mobile terminal to control the transaction of the mobile terminal.

[0041] In the embodiment of the present application, as shown in Figure 2 Step 103 classifies the reference customers to obtain a plurality of reference customer categories, including:

[0042] Step 201: for each reference customer, determining the transaction attribute of the customer according to the transaction data of the customer in a network signal state, and determining the customer attribute of the customer according to the customer information of the customer in the bank;

[0043] Step 202: determining the transaction distance function and the customer distance function according to the transaction attribute and the customer attribute, wherein the independent variable of the two functions is any two customers, and the corresponding function values are the transaction distance and the customer distance of the two customers, respectively;

[0044] Step 203: clustering the reference customers according to the transaction distance function and the customer distance function to obtain a plurality of reference customer categories.

[0045] In an implementation, the clustering of the reference customers according to the transaction distance function and the customer distance function to obtain a plurality of reference customer categories includes:

[0046] Based on the transaction distance function, a clustering algorithm is selected to cluster the reference customers to obtain a plurality of reference customer categories;

[0047] For each reference customer category obtained above, it is determined whether the consistency index of the reference customer category is greater than a set threshold value, and if the consistency index is less than the set threshold value, a clustering algorithm is selected based on the customer distance function to cluster the reference customers until the consistency index of each newly generated reference customer category is greater than the set threshold value; wherein the consistency index of each reference customer category is determined as follows: determining the main transaction type of each reference customer of the reference customer category; determining the maximum value of the proportion of the number of reference customers corresponding to each main transaction category in all reference customers of the reference customer category as the consistency index of the reference customer category.

[0048] In an implementation, the clustering of the reference customers according to the transaction distance function and the customer distance function to obtain a plurality of reference customer categories includes:

[0049] According to the transaction distance function and the customer distance function, a first function is determined, wherein the independent variable of the first function is any two customers, and the corresponding function value is the first distance of the two customers;

[0050] selecting a plurality of reference customers as reference category centers, each reference category center corresponding to a reference customer category, and each reference customer category initially containing only the corresponding reference category center;

[0051] repeating the following steps for each reference customer and for each reference customer category until the change values of the transaction attributes and the customer attributes of all reference category centers are less than a change threshold, thereby obtaining a plurality of reference customer categories:

[0052] for each reference customer, performing the following three steps:

[0053] for each reference category center, determining, according to a first function, a first distance between the reference category center and the reference customer as a distance corresponding to the reference category center;

[0054] from all reference category centers, selecting a plurality of reference category centers having the same main transaction type as the reference customer, determining a minimum value of a plurality of first distances corresponding to the selected reference category centers as a distance A of the reference customer, and determining a reference customer category corresponding to a reference category center corresponding to the minimum value as a reference customer category corresponding to the reference customer; and determining a minimum value of a plurality of distances corresponding to the unselected reference category centers as a distance B of the reference customer.

[0055] if the distance A corresponding to the reference customer is less than or equal to the distance B corresponding to the reference customer, or the distance A corresponding to the reference customer is greater than the distance B corresponding to the reference customer and a difference between the distance A corresponding to the reference customer and the distance B corresponding to the reference customer is less than a specified threshold, then the reference customer is divided into the reference customer category corresponding to the reference customer; otherwise, a new reference category center is created based on the reference customer, the new reference category center corresponds to a new reference customer category, and an initial element of the new reference customer category only contains the reference customer.

[0056] for each customer category, performing the following steps:

[0057] updating the transaction attributes, the customer attributes, and the main transaction type of the reference category center corresponding to the reference customer category according to the transaction attributes, the customer attributes, and the main transaction type of all reference customers contained in the reference customer category; specifically, for any attribute (each dimension of the transaction attributes or the customer attributes, or the main transaction type), if the attribute is a continuous attribute, updating the value of the attribute of the reference category center corresponding to the reference customer category to be the average of all values of the attribute of all reference customers contained in the reference customer category; if the attribute is a discrete attribute, updating the value of the attribute of the reference category center corresponding to the reference customer category to be the value with the largest quantity among all values of the attribute of all reference customers contained in the reference customer category.

[0058] In particular, the transaction attributes can include the number of each transaction category. The customer attributes can include the risk level, the main risk category, the customer asset category, the asset amount, and the like.

[0059] In the embodiment of the present application, as shown in Figure 3 Step 104 determines the risk type to be controlled of the reference customer category in the transaction data corresponding to the no-network signal based on the transaction data of the reference customer category corresponding to the no-network signal, including:

[0060] Step 301: determining the risk type of the reference customer category corresponding to the no-network signal based on the transaction data of the reference customer category corresponding to the no-network signal;

[0061] Step 302: dividing the transaction data of the reference customer category corresponding to the no-network signal into a plurality of no-network signal sub-data in chronological order, so that the transaction quantity of each no-network signal sub-data is greater than the transaction quantity threshold;

[0062] Step 303: for each no-network signal sub-data, determining the proportion of the transaction data of each risk type of the reference customer category corresponding to the no-network signal in the no-network signal sub-data, and determining the proportion as the probability of the no-network signal sub-data corresponding to each risk type;

[0063] Step 304: for each risk type of the reference customer category corresponding to the no-network signal, determining the probability of the reference customer category corresponding to the risk type as the mean of the probabilities of all no-network signal sub-data corresponding to the risk type, and determining the variance of the reference customer category corresponding to the risk type based on the probabilities of all no-network signal sub-data corresponding to the risk type;

[0064] Step 305: determining the lower bound of the probability of the reference customer category corresponding to the risk type as: wherein n is the number of no-network signal sub-data, and ε is the selected probability error threshold;

[0065] Step 306: if the probability of the reference customer category corresponding to the risk type is less than the probability threshold, and the lower bound of the probability of the reference customer category corresponding to the risk type is greater than the lower bound threshold, determining the risk type as the risk type to be controlled of the reference customer category in the no-network signal.

[0066] It should be noted that the lower bound threshold of the probability represents the probability lower bound that the acceptable probability error is less than ε.

[0067] In the embodiment of the present application, as shown in Figure 4 Step 105 determines the risk control model corresponding to the risk type to be controlled based on the transaction data of the reference customer category corresponding to the no-network signal, including:

[0068] Step 401: using the supervised identification of the risk type to be controlled, training the reference customer category on the transaction data corresponding to no network signal by using the machine learning model, and obtaining the risk prediction model corresponding to the risk type to be controlled, wherein the input of the risk prediction model is the transaction data of the customer, and the output is the probability of the transaction data occurring the risk type to be controlled;

[0069] Step 402: determining the control method corresponding to the risk type to be controlled according to the corresponding relationship between the risk type and the control method.

[0070] Specifically, different risk types correspond to different control methods, and the control method can effectively control the occurrence of the risk type. For example, for different risk types, different face recognition thresholds are set, and the face recognition threshold corresponding to each risk type can be set to a certain threshold, so that in the bank customer transaction data meeting the threshold, the risk probability corresponding to the risk type is less than the corresponding threshold, and the threshold is calculated based on the transaction data of all customers of the bank.

[0071] In the embodiment of the application, as shown in Figure 5 Step 106, for a customer only including transaction data corresponding to network signal, determining the reference customer category corresponding to the customer, comprising:

[0072] Step 501: for the customer, determining the transaction attribute of the customer according to the transaction data of the customer corresponding to the network signal, and determining the customer attribute of the customer according to the customer information of the customer in the bank;

[0073] Step 502: for each reference customer, determining the transaction distance and the customer distance between the customer and the reference customer according to the transaction distance function and the customer distance function, determining the transaction distance corresponding to the reference customer as the transaction distance of the reference customer, and determining the customer distance corresponding to the reference customer as the customer distance of the reference customer;

[0074] Step 503: determining the partial order of the reference customer according to the transaction distance and the customer distance, wherein, for any two reference customers, the partial order can be used to determine whether the first reference customer in the two reference customers is closer to the second reference customer;

[0075] Step 504: determining a plurality of maximal reference customers in all reference customers according to the partial order of the reference customer, wherein the maximal reference customer is a maximal element in all reference customers;

[0076] Step 505: determining the plurality of reference customer categories to which the plurality of maximal reference customers belong as the reference customer category corresponding to the customer.

[0077] In the embodiment of the present application, the step 503 determines the partial order of the reference customers according to the transaction distance and the customer distance, comprising:

[0078] For any two reference customers, if the transaction distance corresponding to the first reference customer of the two reference customers is less than or equal to the transaction distance corresponding to the second reference customer of the two reference customers, and the customer distance corresponding to the first reference customer is less than or equal to the customer distance corresponding to the second reference customer, it is determined that the first reference customer is closer to the second reference customer.

[0079] The maximal element of the partial order is that there is no other element superior to the maximal element in the set corresponding to the partial order. The maximal reference customer among all the reference customers is that there is no other reference customer closer to the maximal reference customer among all the reference customers.

[0080] In an embodiment, the step 504 determines a plurality of maximal reference customers among all the reference customers according to the partial order of the reference customers, comprising:

[0081] The maximal authentication value corresponding to each reference customer is initialized as possible, and the corresponding comparison Boolean value is initialized as yes;

[0082] The following steps are performed on each reference customer in turn:

[0083] If the maximal authentication value corresponding to the reference customer is equal to possible, the comparable reference customers corresponding to the reference customer are set as the other reference customers (except the reference customer) among all the reference customers whose corresponding comparison Boolean value is yes; otherwise, the comparable reference customers corresponding to the reference customer are set as null;

[0084] The partial order relationship between the reference customer and each of the corresponding comparable reference customers is determined in turn: if the comparable reference customer is closer to the reference customer, the maximal authentication value corresponding to the reference customer is updated as no; if the reference customer is closer to the comparable reference customer, the maximal authentication value corresponding to the comparable reference customer is updated as no, and the comparable reference customer is determined as the secondary reference customer of the reference customer;

[0085] If it is determined that none of the comparable reference customers corresponding to the reference customer is closer to the reference customer, the reference customer is determined as a maximal reference customer, and the comparison Boolean value of all the secondary reference customers of the maximal reference customer is updated as no.

[0086] In the embodiment of the present application, the step 107 performs risk control on the transaction of the mobile terminal of the customer when the mobile terminal is in a state without network signal, using the risk control model stored in the mobile terminal, comprising:

[0087] When the mobile terminal of the customer is in a network signal weak state, the transaction of the mobile terminal is risk predicted by using the risk prediction model stored in the mobile terminal;

[0088] The transaction of the mobile terminal is risk controlled by using the control method corresponding to the risk type with the predicted probability greater than the specified threshold.

[0089] In the embodiment of the application, the customer can set the transaction elements that need to be controlled when the network signal is weak, and the limited range of the transaction elements. Before the predicted time range corresponding to the weak network signal, the transaction elements that need to be controlled and the limited range of the transaction elements are issued to the mobile terminal of the customer. When the customer transacts, the corresponding value of the transaction element that needs to be controlled is determined according to the transaction data, and it is confirmed whether the corresponding value is within the limited range of the transaction element. If the corresponding value of any one of the transaction elements is not within the limited range of the transaction element, the transaction of the customer is rejected. In this way, the occurrence of transaction risk can be avoided to the greatest extent.

[0090] When the customer transacts in the weak network signal, the transaction data is stored in the local database of the mobile terminal, and the hash value corresponding to the transaction data is stored in the blockchain of the mobile terminal. When the network signal becomes good, the mobile terminal uploads the transaction data and the corresponding hash value to the bank server. The bank server compares the hash value of the uploaded transaction data with the uploaded hash value, and processes the transaction data based on the comparison result.

[0091] Before the predicted time period, part of the historical transaction data of the customer is selected from the customer transaction data stored in the bank server by a random algorithm rule, and is issued to the mobile terminal of the customer. In the condition of the weak network signal, the block data as part of the blockchain node of the mobile terminal of the customer. When the network signal of the mobile terminal of the customer becomes good, the transaction data of the weak network signal of the customer is audited according to the historical transaction data of the blockchain node, the transaction data of the weak network signal, and the historical transaction data of the customer stored in the bank server. Because the data of each block of the blockchain depends on the data stored in the previous blockchain, and the selected part of the data is the historical transaction data of the customer stored in the bank server (this part of the data cannot be tampered), the transaction data in the weak network signal can be ensured not to be easily tampered.

[0092] Among them, part of the historical transaction data of the customer is selected from the customer transaction data stored in the bank server by a random algorithm rule, and is issued to the mobile terminal of the customer, including:

[0093] The risk indicators of each transaction data category corresponding to the customer stored in the bank server are determined;

[0094] According to the risk index, determine the probability of each transaction data category corresponding to the customer; wherein, the higher the risk index, the lower the corresponding probability;

[0095] According to the probability, select a transaction data category from all transaction data categories corresponding to the customer, and filter part of historical transaction data of the selected transaction data category from the transaction data of the customer stored in the bank server.

[0096] The embodiment of the application also provides a transaction risk control device, as described in the following embodiment. Since the principle of solving the problem of the device is similar to that of the transaction risk control method, the implementation of the device can be referred to the implementation of the transaction risk control method, and the repeated parts will not be described here.

[0097] Figure 6 The structure diagram of the transaction risk control device in the embodiment of the application is shown in FIG. 1, which comprises: Figure 6

[0098] The data acquisition module 02 is used for acquiring the historical transaction data of the bank customer;

[0099] The data filtering module 04 is used for filtering out the customers including the transaction data corresponding to no network signal and the transaction data corresponding to network signal based on the historical transaction data of the bank customer, and determining the filtered customers as reference customers;

[0100] The reference customer classification module 06 is used for classifying the reference customers to obtain a plurality of reference customer categories;

[0101] The risk type determination module 08 is used for determining, for each reference customer category, the to-be-controlled risk type of the reference customer category in no network signal based on the transaction data corresponding to no network signal of the reference customer category;

[0102] The risk control model determination module 10 is used for determining, for each to-be-controlled risk type of the reference customer category, the risk control model corresponding to the to-be-controlled risk type based on the transaction data corresponding to no network signal of the reference customer category;

[0103] The reference customer classification module is also used for determining the reference customer category corresponding to the customer containing only the transaction data corresponding to network signal;

[0104] The risk control model issuing module 12 is used for issuing the risk control model corresponding to the to-be-controlled risk type of the reference customer category of the customer to the mobile terminal of the customer, and using the risk control model stored in the mobile terminal to control the transaction of the mobile terminal when the mobile terminal of the customer is in no network signal state.

[0105] ​In the embodiments of the present application, the reference customer classification module 06 is specifically configured to:

[0106] For each reference customer, the transaction attribute of the customer is determined according to the transaction data corresponding to the network signal of the customer, and the customer attribute of the customer is determined according to the customer information of the customer in the bank;

[0107] The transaction distance function and the customer distance function are determined according to the transaction attribute and the customer attribute, wherein the independent variables of the two functions are any two customers, and the corresponding function values are the transaction distance and the customer distance of the two customers, respectively;

[0108] The reference customers are clustered according to the transaction distance function and the customer distance function, and a plurality of reference customer categories are obtained.

[0109] In the embodiments of the present application, the risk type determination module 08 is specifically configured to:

[0110] The risk type corresponding to the network signal of the reference customer category is determined based on the transaction data corresponding to the network signal of the reference customer category;

[0111] The transaction data corresponding to the network signal of the reference customer category is divided into a plurality of network signal-free sub-data in chronological order, so that the transaction quantity of each network signal-free sub-data is greater than the transaction amount threshold;

[0112] For each network signal-free sub-data, the proportion of the transaction data of each risk type corresponding to the network signal of the reference customer category in the network signal-free sub-data is determined, and the proportion is determined as the probability of the network signal-free sub-data corresponding to each risk type;

[0113] For each risk type corresponding to the network signal of the reference customer category, the probability of the reference customer category corresponding to the risk type is determined as the mean of the probabilities of all network signal-free sub-data corresponding to the risk type, and the variance of the reference customer category corresponding to the risk type is determined based on the probabilities of all network signal-free sub-data corresponding to the risk type;

[0114] The lower bound value of the probability of the reference customer category corresponding to the risk type is determined as: Wherein, n is the number of network signal-free sub-data, and ε is the selected probability error threshold;

[0115] If the probability of the reference customer category corresponding to the risk type is less than the probability threshold, and the lower bound value of the probability of the reference customer category corresponding to the risk type is greater than the lower bound threshold of the probability, the risk type is determined as the to-be-controlled risk type of the reference customer category in the network signal.

[0116] In the embodiment of the present application, the risk control model determination module 10 is specifically configured to:

[0117] The risk prediction model corresponding to the to-be-controlled risk type is obtained by training the transaction data corresponding to the network signal of the reference customer category by using the machine learning model, wherein the input of the risk prediction model is the transaction data of the customer, and the output is the probability of the transaction data occurring the to-be-controlled risk type;

[0118] According to the corresponding relationship between the risk type and the control method, the control method corresponding to the to-be-controlled risk type is determined.

[0119] In the embodiment of the present application, the reference customer classification module 06 is further configured to:

[0120] For the customer, the transaction attribute of the customer is determined according to the transaction data corresponding to the network signal of the customer, and the customer attribute of the customer is determined according to the customer information of the customer in the bank;

[0121] For each reference customer, the transaction distance and the customer distance between the customer and the reference customer are determined according to the transaction distance function and the customer distance function, the transaction distance is determined as the transaction distance corresponding to the reference customer, and the customer distance is determined as the customer distance corresponding to the reference customer;

[0122] According to the transaction distance and the customer distance, the partial order of the reference customer is determined, wherein, for any two reference customers, the partial order can be used to determine whether the first reference customer in the two reference customers is closer to the second reference customer;

[0123] According to the partial order of the reference customer, a plurality of maximal reference customers in all reference customers are determined, wherein the maximal reference customer is a maximal element in all reference customers;

[0124] The plurality of reference customer categories to which the plurality of maximal reference customers belong are determined as the reference customer categories corresponding to the customer.

[0125] In the embodiment of the present application, the reference customer classification module 06 is specifically configured to:

[0126] For any two reference customers, if the transaction distance corresponding to the first reference customer in the two reference customers is less than or equal to the transaction distance corresponding to the second reference customer in the two reference customers, and the customer distance corresponding to the first reference customer is less than or equal to the customer distance corresponding to the second reference customer, it is determined that the first reference customer is closer to the second reference customer.

[0127] In the embodiment of the present application, when the mobile terminal of the customer is in a network signal-free state, the transaction of the mobile terminal is controlled by using the risk control model stored in the mobile terminal, comprising:

[0128] when the mobile terminal of the customer is in a no-network signal state, using a risk prediction model stored in the mobile terminal to perform risk prediction on a transaction of the mobile terminal;

[0129] using a control method corresponding to a risk type with a predicted probability greater than a specified threshold to perform risk control on the transaction of the mobile terminal.

[0130] The embodiment of the present application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the transaction risk control method when executing the computer program.

[0131] The embodiment of the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executable on a processor to implement the transaction risk control method.

[0132] The embodiment of the present application also provides a computer program product, which includes a computer program, and the computer program is executable on a processor to implement the transaction risk control method.

[0133] Compared with the prior art, in the embodiment of the present application, the historical transaction data of a bank customer is acquired; based on the historical transaction data of the bank customer, a customer including transaction data corresponding to no network signal and transaction data corresponding to network signal is screened out, and the screened-out customer is determined as a reference customer; the reference customer is classified to obtain multiple reference customer categories; for each reference customer category, based on the transaction data corresponding to no network signal of the reference customer category, a to-be-controlled risk type of the reference customer category in no network signal is determined; for each to-be-controlled risk type of the reference customer category, based on the transaction data corresponding to no network signal of the reference customer category, a risk control model corresponding to the to-be-controlled risk type is determined; for a customer including only transaction data corresponding to network signal, a reference customer category corresponding to the customer is determined; the risk control model corresponding to the to-be-controlled risk type of the reference customer category of the customer is issued to a mobile terminal of the customer, and when the mobile terminal of the customer is in a no-network signal state, the risk control model stored in the mobile terminal is used to perform risk control on a transaction of the mobile terminal, so that the transaction can be controlled in the no-network signal state.

[0134] Those skilled in the art will appreciate that embodiments of the application can be devised for a method, a system, or a computer program product. Accordingly, the present application can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0135] The present application is described in reference to the flowchart and / or block diagrams of the method, apparatus (system) and computer program product according to embodiments of the application. It will be understood that each block of the flowchart and / or block diagrams, and combinations of blocks in the flowchart and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing device 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 and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0136] These computer program instructions can 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 instructions which implement the function specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0137] These computer program instructions can 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 and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0138] The specific embodiments described above are examples for purposes of explanation and illustration and are not intended to limit the scope of the application. The scope of the application is limited only by the claims that follow.

Claims

1. A transaction risk control method, characterized by, The method comprises the following steps: obtaining historical transaction data of bank customers; based on the historical transaction data of the bank customers, screening out customers including transaction data corresponding to no network signal and transaction data corresponding to network signal, and determining the screened-out customers as reference customers; classifying the reference customers to obtain multiple reference customer categories; for each reference customer category, determining the types of risks to be controlled of the reference customer category in the no network signal based on the transaction data corresponding to no network signal of the reference customer category; for each type of risk to be controlled of the reference customer category, determining the risk control model corresponding to the type of risk to be controlled based on the transaction data corresponding to no network signal of the reference customer category; for a customer including only transaction data corresponding to network signal, determining the reference customer category corresponding to the customer; issuing the risk control model corresponding to the type of risk to be controlled of the reference customer category corresponding to the customer to the mobile terminal of the customer, and using the risk control model stored in the mobile terminal to control the transaction of the mobile terminal when the mobile terminal is in a no network signal state.

2. The transaction risk control method of claim 1, wherein, The classification of the reference customers to obtain multiple reference customer categories comprises the following steps: for each reference customer, determining the transaction attribute of the customer according to the transaction data corresponding to network signal of the customer, and determining the customer attribute of the customer according to the customer information of the customer in the bank; determining the transaction distance function and the customer distance function according to the transaction attribute and the customer attribute, wherein the independent variables of the two functions are any two customers, and the corresponding function values are the transaction distance and the customer distance of the two customers, respectively; clustering the reference customers according to the transaction distance function and the customer distance function to obtain multiple reference customer categories.

3. The transaction risk control method of claim 1, wherein, For each reference customer category, determining the types of risks to be controlled of the reference customer category in the no network signal based on the transaction data corresponding to no network signal of the reference customer category comprises the following steps: determining the types of risks corresponding to no network signal of the reference customer category based on the transaction data corresponding to no network signal of the reference customer category; dividing the transaction data corresponding to no network signal of the reference customer category into multiple no network signal sub-data in chronological order, so that the transaction quantity of each no network signal sub-data is greater than a transaction quantity threshold; for each no network signal sub-data, determining the proportion of the transaction data of each type of risk corresponding to no network signal of the reference customer category in the no network signal sub-data, and determining the proportion as the probability of the reference customer category corresponding to each type of risk; for each type of risk corresponding to no network signal of the reference customer category, determining the probability of the reference customer category corresponding to the type of risk as the mean of the probabilities of all no network signal sub-data corresponding to the type of risk, and determining the variance of the reference customer category corresponding to the type of risk based on the probabilities of all no network signal sub-data corresponding to the type of risk. The probability lower bound value of the reference customer category corresponding to the risk type is determined as: Wherein, n is the number of sub-data without network signal, and ε is the selected probability error threshold. If the probability of the reference customer category for the risk type is less than the probability threshold value, and the lower bound of the probability of the reference customer category for the risk type is greater than the lower bound of the probability threshold value, it is determined that the risk type is a risk type to be controlled for the reference customer category in the state of no network signal.

4. The transaction risk control method of claim 1, wherein, For each risk type to be controlled of the reference customer category, based on the transaction data corresponding to the state of no network signal of the reference customer category, a risk control model corresponding to the risk type to be controlled is determined, including: Taking the risk type to be controlled as a supervision identifier, the transaction data corresponding to the state of no network signal of the reference customer category is trained by using a machine learning model to obtain a risk prediction model corresponding to the risk type to be controlled, wherein the input of the risk prediction model is the transaction data of the customer, and the output is the probability of the transaction data occurring the risk type to be controlled; According to the corresponding relationship between the risk type and the control method, the control method corresponding to the risk type to be controlled is determined.

5. The transaction risk control method of claim 2, wherein, For a customer only including transaction data corresponding to the state of network signal, a reference customer category corresponding to the customer is determined, including: For the customer, the transaction attribute of the customer is determined according to the transaction data corresponding to the state of network signal of the customer, and the customer attribute of the customer is determined according to the customer information of the customer in the bank; For each reference customer, the transaction distance and the customer distance between the customer and the reference customer are determined according to the transaction distance function and the customer distance function, the transaction distance is determined as the transaction distance corresponding to the reference customer, and the customer distance is determined as the customer distance corresponding to the reference customer; According to the transaction distance and the customer distance, the partial order of the reference customers is determined, wherein, for any two reference customers, the partial order can be used to determine whether a first reference customer in the two reference customers is closer to a second reference customer in the two reference customers; According to the partial order of the reference customers, a plurality of maximal reference customers in all reference customers are determined, wherein the maximal reference customer is a maximal element in all reference customers; The plurality of reference customer categories to which the plurality of maximal reference customers belong are determined as the reference customer categories corresponding to the customer.

6. The transaction risk control method of claim 5, wherein, According to the transaction distance and the customer distance, the partial order of the reference customers is determined, including: For any two reference customers, if the transaction distance corresponding to a first reference customer in the two reference customers is less than or equal to the transaction distance corresponding to a second reference customer in the two reference customers, and the customer distance corresponding to the first reference customer is less than or equal to the customer distance corresponding to the second reference customer, it is determined that the first reference customer is closer to the second reference customer.

7. The transaction risk control method of claim 1, wherein, When the mobile terminal of the customer is in the state of no network signal, the transaction of the mobile terminal is controlled by using the risk control model stored in the mobile terminal, including: When the mobile terminal of the customer is in the state of no network signal, the transaction of the mobile terminal is controlled by using the risk prediction model stored in the mobile terminal; The transaction of the mobile terminal is controlled by using the control method corresponding to the risk type with the predicted probability greater than a specified threshold value.

8. A transaction risk control apparatus, characterized by comprising: Including: A data acquisition module is configured to acquire historical transaction data of a bank customer; The data screening module is configured to screen out, based on historical transaction data of the bank customers, customers including transaction data corresponding to no network signal and transaction data corresponding to network signal, and determine the screened-out customers as reference customers; The reference customer classification module is configured to classify the reference customers to obtain a plurality of reference customer categories; The risk type determination module is configured to, for each reference customer category, determine, based on the transaction data corresponding to no network signal of the reference customer category, a to-be-controlled risk type of the reference customer category in the no network signal state; The risk control model determination module is configured to, for each to-be-controlled risk type of the reference customer category, determine, based on the transaction data corresponding to no network signal of the reference customer category, a risk control model corresponding to the to-be-controlled risk type; The reference customer classification module is further configured to, for a customer including only transaction data corresponding to network signal, determine a reference customer category corresponding to the customer; The risk control model issuing module is configured to issue, to a mobile terminal of the customer, a risk control model corresponding to a to-be-controlled risk type of a reference customer category corresponding to the customer, and use the risk control model stored in the mobile terminal to control the transaction of the mobile terminal in the no network signal state.

9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the transaction risk control method in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the transaction risk control method in any one of claims 1 to 7.

11. A computer program product, characterised in that, The computer program product includes a computer program, and the computer program is executed by the processor to implement the transaction risk control method in any one of claims 1 to 7.

Images