Digital wallet risk account identification method and system, storage medium and electronic equipment

Abstract

The invention provides a digital wallet risk account identification method and system, a storage medium and electronic equipment, and the method comprises the steps: firstly determining a final risk value of a digital wallet account based on the name of the digital wallet account, a recharging behavior, a withdrawal behavior, a transfer receiving behavior, a login behavior, the number of marked times and an abnormal list screening result of a digital wallet; then determining a risk level of the digital wallet account based on the final risk value; and performing transaction risk control on the digital wallet account based on the risk level. According to the method and the device, the final risk value of the digital wallet is calculated by integrating the wallet attributes of the digital wallet account, and then the risk level of the digital wallet account is determined based on the risk value, so that risk control is performed on the digital wallet account according to the risk level of the digital wallet account accurately, and the security of the digital wallet account is improved.

Description

Technical Field

[0001] The present application relates to the field of financial account risk identification and control technology, and in particular to a digital wallet risk account identification method, system, storage medium and electronic device. Background Art

[0002] Financial accounts are the basis for users to conduct transactions, such as receiving wages from the company, paying or receiving transaction funds, etc. Traditional financial accounts such as users' bank card accounts. With the development of technology, new financial account digital wallets have begun to emerge, but also brought new challenges - risk identification and control of digital wallets.

[0003] Although the risk identification method of traditional bank card accounts can be used to identify and control the risks of digital wallet accounts, the risk identification accuracy of digital wallets is not enough because traditional bank card accounts do not have the wallet attributes of digital wallets. At the same time, the risk types of digital wallets are different from those of traditional bank card accounts, which also brings challenges to the risk identification and control of digital wallet accounts.

[0004] Therefore, the accuracy of the existing risk identification method for digital wallet accounts is relatively low, and it is unable to effectively protect the security of the user's digital wallet account. There is an urgent need for a method that can identify and control risks in digital wallet accounts to improve the accuracy of risk identification of digital wallet accounts, thereby improving the security of the user's digital wallet account. Summary of the invention

[0005] In view of this, the present application provides a digital wallet risk account identification method, system, storage medium and electronic device to solve or partially solve the problem that the existing risk identification method for digital wallet accounts has a relatively low accuracy rate and cannot effectively protect the security of users' digital wallet accounts, thereby improving the accuracy rate of risk identification of digital wallet accounts and thus improving the security of users' digital wallet accounts.

[0006] To achieve the above objectives, this application provides the following technical solutions:

[0007] The first aspect of the present application discloses a method for identifying risky accounts in a digital wallet, comprising:

[0008] Determine the final risk value of the digital wallet account based on the name of the digital wallet account, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked, and the abnormal list screening result of the digital wallet;

[0009] Determining a risk level of the digital wallet account based on the final risk value;

[0010] Perform transaction risk control on the digital wallet account based on the risk level.

[0011] Optionally, determining the final risk value of the digital wallet based on the name of the digital wallet, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked, and abnormal list screening result of the digital wallet includes:

[0012] Determining a sensitivity score based on the name of the digital wallet;

[0013] Determining an abnormal recharge score of the digital wallet based on the recharge behavior of the digital wallet account within a first preset time period;

[0014] Determining an abnormal withdrawal score of the digital wallet based on the withdrawal behavior of the digital wallet account within a second preset time period;

[0015] Determining an abnormal transfer score of the digital wallet based on the transfer receiving behavior of the digital wallet account within a third preset time period;

[0016] Determining an abnormal login score of the digital wallet account based on the login behavior of the digital wallet account on an uncommon device;

[0017] Determining a greylist screening score for the digital wallet account based on the number of times the digital wallet is marked;

[0018] Determining an abnormal list screening score for the digital wallet account based on the abnormal list screening result;

[0019] A final risk value is determined based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score.

[0020] Optionally, performing transaction risk control on the digital wallet account based on the risk level includes:

[0021] If a transaction is detected in the digital wallet account, determining a transaction limit of the digital wallet account based on the risk level and the transaction type of the transaction;

[0022] Determine whether the transaction meets the transaction limit, if so, allow the transaction, if not, block the transaction.

[0023] Optionally, the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the grey list screening score and the abnormal list screening score determine a final risk value, including:

[0024] Determine the weights corresponding to the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score, and the abnormal list screening score respectively; wherein the sum of all weights is 1;

[0025] The sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score are weighted and summed based on various weights to obtain the final risk value of the digital wallet account.

[0026] Optionally, the process of determining each weight includes:

[0027] Determine various weight adjustment parameters;

[0028] For each of the weight adjustment parameters, determining the classification accuracy and weight sequence corresponding to the weight adjustment parameter based on the weight introspection learning method; the weight sequence includes the respective weights;

[0029] The maximum classification accuracy is determined from the classification accuracy and weight sequences corresponding to the weight adjustment parameters of the respective weights, and the weight sequence corresponding to the maximum classification accuracy is used as the final weight sequence.

[0030] The second aspect of the present application discloses a digital wallet risk account identification system, comprising:

[0031] A risk value determination unit, configured to determine a final risk value of the digital wallet account based on the name of the digital wallet account, the recharge behavior, the withdrawal behavior, the transfer receiving behavior, the login behavior, the number of times it is marked, and the abnormal list screening result of the digital wallet;

[0032] a risk level determination unit, configured to determine a risk level of the digital wallet account based on the final risk value;

[0033] A risk control unit performs transaction risk control on the digital wallet account based on the risk level.

[0034] Optionally, the risk value determination unit includes:

[0035] A sensitivity score module for determining a sensitivity score based on a name of a digital wallet;

[0036] An abnormal recharge scoring module, used to determine an abnormal recharge score of the digital wallet based on the recharge behavior of the digital wallet account within a first preset time period;

[0037] An abnormal withdrawal score module, used to determine the abnormal withdrawal score of the digital wallet based on the withdrawal behavior of the digital wallet account within a second preset time period;

[0038] an abnormal transfer score module, configured to determine an abnormal transfer score of the digital wallet based on the transfer receiving behavior of the digital wallet account within a third preset time period;

[0039] an abnormal login score module, used to determine an abnormal login score of the digital wallet account based on the login behavior of the digital wallet account on an uncommon device;

[0040] A greylist screening score module, which determines a greylist screening score of the digital wallet account based on the number of times the digital wallet is marked;

[0041] An abnormal list screening score module, used to determine the abnormal list screening score of the digital wallet account based on the abnormal list screening result;

[0042] A risk value determination module is used to determine a final risk value based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score.

[0043] Optionally, the risk control unit includes:

[0044] a transaction type determination module, configured to determine a transaction limit of the digital wallet account based on the risk level and the transaction type of the transaction when a transaction is detected in the digital wallet account;

[0045] A determination module, used to determine whether the transaction meets the transaction limit;

[0046] The transaction control module is used to allow the transaction when the transaction meets the transaction limit, and to block the transaction when the transaction does not meet the transaction limit.

[0047] The third aspect of the present application discloses an electronic device, comprising:

[0048] one or more processors;

[0049] a storage device having one or more programs stored thereon;

[0050] When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 5.

[0051] A fourth aspect of the present application discloses a computer storage medium having a computer program stored thereon, wherein the computer program implements the method according to any one of claims 1 to 5 when executed by a processor.

[0052] It can be seen from the above technical solutions that in a digital wallet risk account identification method, system, storage medium and electronic device provided by the present application, the final risk value of the digital wallet account is first determined based on the name of the digital wallet account, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked and the abnormal list screening result of the digital wallet; then the risk level of the digital wallet account is determined based on the final risk value; and transaction risk control is performed on the digital wallet account based on the risk level. The present application calculates the final risk value of the digital wallet by comprehensively analyzing the various wallet attributes of the digital wallet account, and then determines the risk level of the digital wallet account based on the risk value, thereby accurately controlling the risk of the digital wallet account based on the risk level of the digital wallet account, thereby improving the security of the digital wallet account. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0054] Figure 1 A flowchart of a method for identifying risky accounts of a digital wallet disclosed in an embodiment of the present application;

[0055] Figure 2 This is a flowchart of an implementation of step S130 disclosed in another embodiment of the present application;

[0056] Figure 3 A flowchart of weight adjustment disclosed in another embodiment of the present application;

[0057] Figure 4 A flow chart of risk control disclosed in another embodiment of the present application;

[0058] Figure 5 A schematic diagram of a digital wallet risk account identification system disclosed in another embodiment of the present application;

[0059] Figure 6 A schematic diagram of an electronic device disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0061] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0062] Furthermore, in this document, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0063] This application needs to solve the problem that the accuracy of the risk identification method for digital wallet accounts is relatively low and cannot effectively protect the security of users' digital wallet accounts, so as to improve the security of users' digital wallet accounts.

[0064] Digital wallets are a financial service provided based on mobile phone applications. The service mainly involves two roles: customers and financial institutions. Before using digital wallets for financial transactions, customers need to complete two major processes: wallet opening and real-name authentication. When opening a wallet, customers need to set a wallet nickname, which can be modified at any time after opening. During real-name authentication, customers need to enter their ID cards for verification. After real-name authentication, customers can transfer and consume. After binding a bank card, customers can make top-up and cash withdrawal transactions. At the same time, applications that provide digital wallets can pull up other financial payment tools on the device to recharge the digital wallet with program authorization.

[0065] The applicant found through research that fraudsters used the above-mentioned digital wallet usage characteristics to defraud other customers' funds mainly in the following ways:

[0066] (1) Fraudsters impersonate financial institutions / investment institutions / regulatory agencies, etc. in their nicknames, deceiving customers to transfer money from their wallets to fraudulent wallets;

[0067] (2) Fraudsters use various rebate activities as an excuse to induce customers to open wallets and obtain customer and account information, illegally log into customer accounts to perform top-up and transfer operations, and transfer customer funds to fraudulent wallets;

[0068] (3) Induce customers to log into the fraudulent wallet account on their own devices and authorize the digital wallet program to pull up other financial payment tools on the device to recharge the digital wallet.

[0069] Fraudulent behavior causes the digital wallet used for fraud to have more income records in a short period of time. At the same time, the fraudsters will quickly withdraw the money in the fraudulent wallet to the bank card, and then transfer the fraudulent funds from the bank card.

[0070] Therefore, this application combines the user's digital wallet nickname, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, the number of times the digital wallet is marked, and the abnormal list screening results to comprehensively determine the security of the digital wallet account, and then determine the risk level of the digital wallet account, and then perform risk control on each transaction of the digital wallet based on the risk level. Thereby improving the security of the user's digital wallet account.

[0071] This application provides a method for identifying risky accounts in a digital wallet, such as Figure 1 As shown, the following steps are included:

[0072] S110. Determine a final risk value of the digital wallet account based on the name of the digital wallet account, the recharge behavior, the withdrawal behavior, the transfer receiving behavior, the login behavior, the number of times it is marked, and the abnormal list screening result of the digital wallet account.

[0073] In this application, the risk value of the digital wallet account is comprehensively determined by combining the name of the digital wallet account, the account's recharge behavior, withdrawal behavior, and transfer behavior, the number of times it is marked, and the abnormal list screening results of the digital wallet account. It should be noted that the name of the digital wallet account can be a nickname given by the user to the digital wallet account.

[0074] In another embodiment, step S110 in the present application determines the final risk value of the digital wallet based on the name of the digital wallet, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked and the abnormal list screening result of the digital wallet, specifically including steps S111-S118.

[0075] S111. Determine a sensitivity score based on the name of the digital wallet.

[0076] The sensitivity score refers to the similarity score between the digital wallet nickname and the sensitive words in the sensitive word library. At the same time, since the nickname can be modified multiple times, the modification frequency is also used as a calculation factor when calculating the score. The higher the score, the greater the probability that the wallet is impersonating a financial institution / investment institution / regulatory agency, etc. The sensitivity score S1 calculation formula is as follows:

[0077]

[0078] Among them, P 1 ~P n is the similarity between the nickname of the digital wallet after a single modification and the sensitive word, F 1 The number of times the digital wallet nickname has been modified.

[0079] The similarity between nickname and sensitive words P i (P 1 ~P n ) The calculation method is: calculate the similarity SPi of the nickname of the digital wallet after a single modification with each sensitive word in the sensitive word library (the sensitive word library covers major official institutions) in turn, and take the maximum value of the similarity SPi as the value of Pi.

[0080] Among them, for each similarity SPi, the dynamic programming method is used to calculate the similarity SPi between a single sensitive word and the wallet nickname, which mainly includes the following steps:

[0081] (1) Initialize an invalid character library, such as spaces, *, #, @, etc. Before detecting the digital wallet nickname, remove the invalid characters in the nickname text Vi;

[0082] Since the financial institutions / investment institutions / regulatory agencies impersonated by the scammers generally do not contain invalid characters in their names, the similarity between the scammer accounts and the official agency names can be further increased by removing invalid characters from the digital wallet account names.

[0083] (2) Assume that the length of a single sensitive word text Ui in the sensitive word library is Ki, create a two-dimensional array of the size of the two text lengths Vi and Ui, and compare whether the characters in the corresponding rows and columns of each point in the array are equal. If they are equal, set them to 1, and if they are not equal, set them to 0. Then the slashes formed by the numbers with a value of 1 in the array are equal substrings. The longest slash is the longest substring; let the length of the longest string be Li. The similarity Spi is calculated based on the following formula:

[0084]

[0085] S112, determining an abnormal recharge score of the digital wallet based on the recharge behavior of the digital wallet account during the first preset period of time;

[0086] Abnormal recharge refers to the digital wallet being 1 Within 10 days, recharge transactions are made on non-common devices, and the transaction amount is greater than or equal to M 3 For transaction behavior, abnormal recharge score S 3 The calculation formula is as follows:

[0087]

[0088] Among them, F 31 The number of abnormal recharges for the digital wallet, F 30 For this digital wallet in D 3 The number of all recharge transactions within the time period. 1 The duration is set according to actual needs. 3 A first amount threshold value for determining whether each recharge is an abnormal recharge is set based on actual needs.

[0089] S113, determining an abnormal withdrawal score of the digital wallet based on the withdrawal behavior of the digital wallet account within a second preset time period;

[0090] Abnormal withdrawal refers to the second preset time T 2 The amount of money carried out is greater than or equal to M 4 The withdrawal transaction behavior is S. 4 The calculation formula is as follows:

[0091]

[0092] Among them, F 41 The number of abnormal withdrawals from the digital wallet, F 40 For this digital wallet in D 4 The number of all withdrawal transactions within the time period. 4 A second amount threshold value is set based on actual needs to determine whether each withdrawal is an abnormal withdrawal.

[0093] It should be noted that when determining the abnormal withdrawal score of a digital wallet account, the number of bank cards bound to the digital wallet can be further audited for calculation. 4 The digital wallet of Zhang directly determines that the abnormal withdrawal score of the account is 0; and for the number of bank cards bound to no less than C 4 The digital wallet of the card holder calculates the abnormal withdrawal score based on the above formula. Alternatively, the bank card number weight W is determined based on the number of bank cards bound to the digital wallet. 0 , then the above abnormal withdrawal score S 4 The calculation formula is changed to:

[0094]

[0095] It should also be noted that the fewer the number of bank cards bound to the digital wallet, the greater the weight of the number of bank cards W. 0 The smaller the value of; or, the number of bound bank cards is not less than C 4 Zhang's digital wallet setting W 0 The value of 1, and the number of bound bank cards is less than C 4 Zhang's digital wallet setting W 0 The value of is less than 1.

[0096] Among them, C 4 It can be set according to actual needs, for example, to 3 or 5 sheets, or other numbers.

[0097] The second preset duration T 2 Can be combined with the first preset time T 1 The same, or different.

[0098] S114, determining an abnormal transfer score of the digital wallet based on the transfer receiving behavior of the digital wallet account within a third preset time period;

[0099] Abnormal transfer refers to the digital wallet in the third preset time period T 3 During the time period, the income is transferred and the transaction amount is greater than or equal to M 5 Transaction behavior, abnormal transfer score S 5 The calculation formula is as follows:

[0100]

[0101] Among them, F 51 is the number of abnormal transfer income of the digital wallet, F 50 For this digital wallet in T 3 The number of all transfer income transactions within the time period. 5 A third amount threshold value is set based on actual needs to determine whether each abnormal transfer income is abnormal transfer income.

[0102] It should be noted that the third preset time length T 3 Can be combined with the first preset time T 1 and the second preset duration T 2 The same as at least one of the first preset time length T 1 and the second preset duration T 2 Are different.

[0103] S115. Determine an abnormal login score of the digital wallet account based on the login behavior of the digital wallet account on an uncommon device;

[0104] Abnormal login refers to the behavior of logging into the digital wallet on an uncommon device. The abnormal login score is S. 6 The calculation formula is as follows:

[0105]

[0106] Among them, F 61 is the number of abnormal logins to the digital wallet, F 60 is the number of logins to the digital wallet, F 60 Includes normal login times and abnormal login times.

[0107] It should be noted that F 61 With F 60 It should be statistical data within the same time period. It can count all login times since the digital wallet account was created, or it can only count statistical data within the most recent period, such as the most recent year, the most recent six months, or the most recent month.

[0108] S116. Determine a gray list screening score for the digital wallet account based on the number of times the digital wallet is marked;

[0109] The grey list refers to the list of wallet accounts that have been flagged by customer complaints or their channels. Grey list screening score S 7 The calculation formula is as follows:

[0110]

[0111] Among them, F 71 is the number of times the digital wallet is marked, F 70 Accumulate the number of markings for all digital wallets; or F 70 It can also be the maximum number of times all digital wallets have been marked.

[0112] S117. Determine an abnormal list screening score for the digital wallet account based on the abnormal list screening result.

[0113] The abnormal list refers to the list of people who have been certified to have bad trading behavior. Abnormal list screening score S 2 The calculation formula is as follows:

[0114] S 2 =B 2 ×100;

[0115] Among them, B 2 The value is 1 or 0. 1 means that the identity card information of the digital wallet real-name authentication exists in the exception list, that is, the user to which the digital wallet account belongs is a user on the exception list; 0 means that the identity card information of the digital wallet real-name authentication does not exist in the exception list, that is, the user to which the digital wallet account belongs is not a user on the exception list.

[0116] It should be noted that steps S111-S117 in the present application can be executed simultaneously or in other orders, and the above order is only an example.

[0117] S118. Determine a final risk value based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score, and the abnormal list screening score.

[0118] It can be seen that the risk identification of digital wallet accounts in this application starts from the business experience, and integrates the anonymity of wallet nicknames, transaction device attributes, and diverse dimensions in digital wallet transactions, and the identification factors are more comprehensive.

[0119] S120. Determine a risk level of the digital wallet account based on the final risk value.

[0120] If the final risk value is between [85,100], the corresponding digital wallet account is high risk; if the final risk value is between [70,85), the corresponding digital wallet account is high risk; if the final risk value is between [0,70), the corresponding digital wallet account is high risk; the risk classification range can be manually adjusted based on business experience.

[0121] S130: Perform transaction risk control on the digital wallet account based on the risk level.

[0122] In a digital wallet risk account identification method, system, storage medium and electronic device provided by the present application, the final risk value of the digital wallet account is first determined based on the nickname of the digital wallet account, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked and the abnormal list screening result of the digital wallet; then the risk level of the digital wallet account is determined based on the final risk value; and transaction risk control is performed on the digital wallet account based on the risk level. The present application calculates the final risk value of the digital wallet by comprehensively considering the various wallet attributes of the digital wallet account, and then determines the risk level of the digital wallet account based on the risk value, thereby accurately controlling the risk of the digital wallet account based on the risk level of the digital wallet account, thereby improving the security of the digital wallet account.

[0123] In another embodiment, step S130 in the present application performs transaction risk control on the digital wallet account based on the risk level, such as Figure 2 As shown, including;

[0124] S131. If a transaction is detected in the digital wallet account, determine a transaction limit of the digital wallet account based on the risk level and the transaction type of the transaction;

[0125] In step S131, for each transaction in the digital wallet account, it is necessary to determine the transaction limit of the transaction separately. The transaction limit is determined based on the risk level of the accounts of both parties to the transaction and the type of transaction; or, the transaction type is determined based on the risk level of the accounts of both parties to the transaction, the account types of both parties to the transaction and the transaction type of the transaction.

[0126] It should be noted that the transaction limit may include one or more of a single transaction limit, a daily cumulative limit, a monthly cumulative limit, and an annual cumulative limit. Each transaction limit may be determined based on the user's annual income data in the previous year.

[0127] It should also be noted that before executing step S131, the process may also include: prompting the transaction parties' risk levels to the transaction customers; reminding the customers whether to continue the transaction, and then executing step S131 if the customers choose to continue the transaction.

[0128] S132: Determine whether the transaction meets the transaction limit. If so, allow the transaction; if not, block the transaction.

[0129] If the transaction amount is lower than the transaction limit, it is determined that the transaction meets the transaction limit; otherwise, it is determined that the transaction amount does not meet the transaction limit.

[0130] If the transaction limit includes a single transaction limit, a daily cumulative limit, a monthly cumulative limit and an annual cumulative limit, it is necessary to determine whether the transaction amount of the transaction is lower than the single transaction limit, whether the transaction amount of the daily cumulative transaction is lower than the daily cumulative limit, whether the transaction amount of the monthly cumulative transaction is lower than the monthly cumulative limit, and whether the transaction amount of the annual cumulative transaction is lower than the annual cumulative limit. If all of them are lower, it is determined that the transaction limit is met. If any one of them is not lower than the corresponding transaction limit, it is determined that the transaction amount of the transaction does not meet the transaction limit.

[0131] In another embodiment, step S118 in the present application determines the final risk value based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score, specifically including:

[0132] S201. Determine the weights corresponding to the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score respectively; wherein the sum of all weights is 1.

[0133] Each score in this application has a corresponding weight, such as: sensitivity score S 1 The corresponding sensitivity score weight W 1, abnormal list screening score S 2 The corresponding abnormal list screening score weight W 2 , abnormal recharge score S 3 The corresponding abnormal recharge score weight W 3 , abnormal withdrawal score S 4 The corresponding abnormal withdrawal score weight W 4 , abnormal transfer score S 5 The corresponding abnormal transfer score weight W 5 , abnormal login score S 6 The corresponding weighted abnormal login score is W 6 , and the greylist screening score S 7 The corresponding gray list screening score weight W 7 .

[0134] It should be noted that the sum of all weights in this application is 1; for example, for the above scenario containing 7 weights corresponding to the scores, W 1 +W 2 +W 3 +W 4 +W 5 +W 6 +W 7 =1.

[0135] S202. Based on various weights, weighted sum is performed on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score, and the abnormal list screening score to obtain the final risk value of the digital wallet account.

[0136] In this application, each score is multiplied by the corresponding weight, and each product is accumulated to obtain the final risk value of the digital wallet, such as by referring to the following formula for calculation:

[0137] S=S 1 ×W 1 +S 2 ×W 2 +S 3 ×W 3 +S 4 ×W 4 +S 5 ×W 5 +S 6 ×W 6 +S 7 ×W 7 ;

[0138] Among them, S represents the final risk value of the digital wallet account.

[0139] In another embodiment, the above-mentioned weights W1 ~W 7 The determination process may include:

[0140] S301: Determine various weight adjustment parameters.

[0141] Different weight adjustment parameters will result in different weight adjustment amplitudes in each iteration of the weight introspection learning method. Therefore, by setting different weight adjustment parameters, the weights can be adjusted with different adjustment amplitudes.

[0142] S302. For each of the weight adjustment parameters, determine the classification accuracy and weight sequence corresponding to the weight adjustment parameter based on the weight introspection learning method; the weight sequence includes the respective weights.

[0143] The weight sequence includes the weights corresponding to the above scores, that is, the weight sequence includes the weight W 1 ~W 7 .

[0144] S303: Determine the maximum classification accuracy from the classification accuracy and weight sequences corresponding to the weight adjustment parameters of the respective weights, and use the weight sequence corresponding to the maximum classification accuracy as the final weight sequence.

[0145] It can be seen that this application introduces a weight adjustment mechanism, which can learn and update each score weight based on actual data from the perspective of data self-learning, so that the weight setting is more reasonable, the data classification accuracy is higher, and the automatic adjustment of parameters is realized.

[0146] Obviously, this application parameterizes the dimensional conditions and weights in the risk identification of digital wallet accounts, which can not only realize automatic adjustment of parameters according to the weight adjustment mechanism, but also manually adjust the parameters in the system based on human experience, thereby improving the flexibility, maintainability and accuracy of the model.

[0147] For example, the weight W 1 ~W 7 The adjustment determination process, such as Figure 3 As shown, it can be:

[0148] In the database, we select the fraud list (verified fraud digital wallets), medium-risk account data, and low-risk account data as the training set and historical database data. The three types of data are evenly distributed in the training set and historical database data. 1 -S 7 ) are used as attributes for matching, and the threshold ε is set to determine whether it is a matching attribute. match =0.1.

[0149] Step 1: Initialize parameters: Set the number of cases in the training set to N tr . Let the number of iterations t = 0, and let each attribute weight W i (i=1,2,…,m) assigns an average weight of 1 / m, where m is the number of attributes and the weight adjustment parameter Δ i =1, calculate the classification accuracy of the training set under the average weight, recorded as ACC tr (t).

[0150] Step 2: Let t = t + 1, sequentially take the t-th case in the training set, and use the nearest neighbor strategy in the prior art to retrieve the most similar neighbor case from the history library. The nearest neighbor case can be one case or multiple cases.

[0151] Step 3: Apply a preset weight update strategy to the nearest neighbor case to learn and update the attribute weight. The preset weight update strategy may be the GMU+GUD weight update strategy in the prior art;

[0152] Among them, the GMU+GUD weight update strategy includes:

[0153] Suppose there are two cases C in the same historical case library I and C II , where case C I Property X I Include X I =(x 1,I ,x 2,I ,...,x i,I ,...,x m,I ), Case C II Property X II Include X II =(x 1,II ,x 2,II ,...,x i,II ,…,x m,II ), if 0≤|x i,I -x i,II |≤ε match , then it is called case C I With Case C II The i-th attribute of is a matching attribute; otherwise, it is called a non-matching attribute. i,I For Case C I The i-th attribute, x i,II For Case C II The i-th attribute of match It is the threshold for determining whether an attribute matches, indicating the closeness of the same attribute of two cases.

[0154] The way to adjust the weight is also a key point in introspective learning. The most basic method is to increase or decrease or multiply or divide a fixed amount. The former ensures that the change of weight is a constant during the learning process; the latter makes the weight decay with the learning process. Optionally, consider the impact of the number of case attributes on the case attribute weight, and use formula (1) and formula (2) to adjust the weight.

[0155] Weight increase:

[0156]

[0157] Weight reduction:

[0158]

[0159] Among them, w i (t) is the weight of the ith attribute at the tth iteration; w i (t+1) The weight of the ith attribute at the t+1th iteration; Δ i / m determines the amount of change in the weight, where m is the number of case attributes. For different classification problems, if the number of attributes is different, even if Δ i The same, the weight change is also different. However, for the same case library, m is fixed, so the weight change is mainly related to Δ i To achieve effective adjustment of attribute weights and ensure that the attribute weight adjustment value does not exceed the sum of the attribute weights, Δ i The value is taken in the interval (0, m), and the updated weight is not less than zero in accordance with formula (1) and formula (2).

[0160] Step 4: Normalize the updated attribute weights to obtain the attribute weights learned in the tth iteration and save them.

[0161] After the weights of all attributes of a case are adjusted, in order to ensure that the sum of the weights of all attributes is 1, the weights of each attribute need to be normalized and the sum of the adjusted weights is readjusted to 1.

[0162] Step 5: Reclassify all cases in the training set according to the new weights of the tth iteration and calculate the new accuracy ACC of the training set tr (t).

[0163] If t <N tr , go to step 2; otherwise, go to step 6.

[0164] Step 6: Store Δ at this time i Take the maximum ACC value tr (t) The corresponding weight W i (i=1,2,…,m).

[0165] Step 7: If Δ i <m,Δ i =Δ i +1, reset the number of iterations t=0 and go to step 2; otherwise, go to step 8.

[0166] Step 8: According to all Δ i Take the maximum accuracy of all training sets under the value, select the weight corresponding to the maximum accuracy as the optimal weight under the current introspective learning strategy, and the introspective learning training process of the weight is completed.

[0167] In another embodiment of the present application, during real-time transactions, the transaction risk control module performs a specific process of transaction risk control on the digital wallet account based on the risk level, such as Figure 4 As shown, it mainly includes the following operations:

[0168] (1) Obtain risk level data from the model results and provide prompts to customers in transactions.

[0169] (2) If the current customer chooses to continue the transaction, the module will read the transaction limit corresponding to the digital wallet of the transaction from the classified risk control table based on the risk level of the digital wallet involved in the transaction. The limit is set from three dimensions: transaction scenario, digital wallet, and bank card, covering recharge, withdrawal, transfer, and consumption transaction scenarios, achieving a more refined risk control effect.

[0170] (3) If the transaction meets the limit, it is allowed to proceed; otherwise, it is terminated.

[0171] The classification risk control table is maintained by the system business administrator, as shown in Classification Risk Control Table 1:

[0172] Table 1

[0173]

[0174] Among them, the risk scores [a,b), [b,c) and [c,d] are the risk scores corresponding to low risk, medium risk and high risk respectively, among which a<b<c≤d. The specific limits in each of the classification risk control tables can be set according to the actual situation, but for the limits in the same row, the setting principle is: single limit ≤ not more than the daily cumulative limit ≤ annual cumulative limit.

[0175] In another embodiment of the present application, steps S110 and S120 may be executed by a risk assessment model. Specifically, the process of the risk assessment model executing step S110 may refer to the above steps S111-S118.

[0176] Among them, the model parameter table of the risk assessment model is shown in Table 2:

[0177] Table 2

[0178] Parameter meaning Parameter name Parameter Value Sensitivity score weight <![CDATA[W 1 ]]> - Abnormal list screening score weight <![CDATA[W 2 ]]> - Abnormal recharge score weight <![CDATA[W 3 ]]> - Abnormal withdrawal score weight <![CDATA[W 4 ]]> - Score weight for abnormal transfers <![CDATA[W 5 ]]> - Abnormal login score weight <![CDATA[W 6 ]]> - Greylist screening score weight <![CDATA[W 7 ]]> - Abnormal recharge-duration <![CDATA[T 1 ]]> - Abnormal recharge amount <![CDATA[M 3 ]]> - Abnormal withdrawal-duration <![CDATA[T 2 ]]> - Abnormal withdrawal amount <![CDATA[M 4 ]]> - Abnormal withdrawal-number of cards <![CDATA[C 4 ]]> - Abnormal transfer-duration <![CDATA[T 3 ]]> - Abnormal transfer amount <![CDATA[M 5 ]]> -

[0179] Accordingly, the risk assessment model can input the assessment results according to the model result table, and the model result table is shown in Table 3:

[0180] Table 3

[0181]

[0182] Among them, the risk assessment model introduces a model adjustment mechanism, which can learn and update the model attribute weights based on actual data from the perspective of data self-learning, so that the model weight setting is more reasonable, the data classification accuracy is higher, and the automatic adjustment of parameters is realized.

[0183] Another embodiment of the present application also discloses a digital wallet risk account identification system, such as Figure 5 As shown, including:

[0184] The risk value determination unit 501 is used to determine the final risk value of the digital wallet account based on the nickname, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked, and the abnormal list screening result of the digital wallet;

[0185] A risk level determination unit 502, configured to determine a risk level of the digital wallet account based on the final risk value;

[0186] The risk control unit 503 performs transaction risk control on the digital wallet account based on the risk level.

[0187] In this embodiment, the specific execution process of the risk value determination unit 501, the risk level determination unit 502 and the risk control unit 503 can be found in the corresponding Figure 1 The content of the method embodiment will not be repeated here.

[0188] In a digital wallet risk account identification system provided by the present application, the final risk value of the digital wallet account is first determined by a risk value determination unit based on the nickname, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked, and abnormal list screening results of the digital wallet; then the risk level of the digital wallet account is determined based on the final risk value by a risk level determination unit; finally, a statistical ticket risk control unit performs transaction risk control on the digital wallet account based on the risk level. The present application calculates the final risk value of the digital wallet by comprehensively analyzing the various wallet attributes of the digital wallet account, and then determines the risk level of the digital wallet account based on the risk value, thereby accurately controlling the risk of the digital wallet account based on the risk level of the digital wallet account, thereby improving the security of the digital wallet account.

[0189] Optionally, the risk value determination unit 501 includes:

[0190] A sensitivity score module for determining a sensitivity score based on a name of a digital wallet;

[0191] An abnormal recharge scoring module, used to determine an abnormal recharge score of the digital wallet based on the recharge behavior of the digital wallet account within a first preset time period;

[0192] An abnormal withdrawal score module, used to determine the abnormal withdrawal score of the digital wallet based on the withdrawal behavior of the digital wallet account within a second preset time period;

[0193] an abnormal transfer score module, configured to determine an abnormal transfer score of the digital wallet based on the transfer receiving behavior of the digital wallet account within a third preset time period;

[0194] an abnormal login score module, used to determine an abnormal login score of the digital wallet account based on the login behavior of the digital wallet account on an uncommon device;

[0195] A greylist screening score module, which determines a greylist screening score of the digital wallet account based on the number of times the digital wallet is marked;

[0196] The abnormal list screening score module is used to determine the abnormal list screening score of the digital wallet account based on the abnormal list screening result.

[0197] A risk value determination module is used to determine a final risk value based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score.

[0198] Among them, the specific implementation process of the abnormal recharge scoring module, the abnormal withdrawal scoring module, the abnormal transfer scoring module, the abnormal login scoring module, the gray list screening scoring module, the abnormal list screening scoring module and the risk value determination module can be found in the method implementation example of steps S111-S118, which will not be repeated here.

[0199] Optionally, the risk control unit 503 includes:

[0200] a transaction type determination module, configured to determine a transaction limit of the digital wallet account based on the risk level and the transaction type of the transaction when a transaction is detected in the digital wallet account;

[0201] A determination module, used to determine whether the transaction meets the transaction limit;

[0202] The transaction control module is used to allow the transaction when the transaction meets the transaction limit, and to block the transaction when the transaction does not meet the transaction limit.

[0203] In this embodiment, the specific execution process of the transaction type determination module, the judgment module and the transaction control module can be found in the corresponding Figure 2 The content of the method embodiment will not be repeated here.

[0204] Optionally, in another embodiment of the present application, the digital wallet risk account identification system may further include: a weight adjustment unit;

[0205] The weight adjustment unit is used to determine each weight adjustment parameter; for each weight adjustment parameter, the classification accuracy and weight sequence corresponding to the weight adjustment parameter are determined based on the weight introspection learning method; the weight sequence includes the various weights; from the classification accuracy and weight sequence corresponding to the weight adjustment parameters of each weight, the maximum classification accuracy is determined, and the weight sequence corresponding to the maximum classification accuracy is used as the final weight sequence.

[0206] The specific implementation process of the weight adjustment unit can refer to the method embodiment content of steps S301-303 and Figure 3 The contents of the corresponding embodiments will not be repeated here.

[0207] The digital wallet risk account identification system in this application is mainly divided into three modules: model operation, model adjustment, and transaction risk control. The linkage between model identification and transaction risk control forms a closed loop of risk control, including pre-warning, in-process blocking, and post-analysis. Model adjustment can timely adjust the weights of various attributes of the model, which can achieve more accurate risk identification and control.

[0208] Another embodiment of the present application further provides an electronic device, such as Figure 6As shown, specifically including:

[0209] One or more processors 601 .

[0210] The storage device 602 stores one or more programs.

[0211] When one or more programs are executed by one or more processors 601 , the one or more processors 601 implement any one of the methods in the above embodiments.

[0212] Another embodiment of the present application further provides a computer storage medium having a computer program stored thereon, wherein the computer program implements any one of the methods in the above embodiments when executed by a processor.

[0213] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0214] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0215] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying risky accounts in a digital wallet, characterized in that: The method comprises: Determine the final risk value of the digital wallet account based on the name of the digital wallet account, recharge behavior, withdrawal behavior, transfer receiving behavior, login behavior, number of times marked, and the abnormal list screening result of the digital wallet; Determining a risk level of the digital wallet account based on the final risk value; Perform transaction risk control on the digital wallet account based on the risk level.

2. The method according to claim 1, characterized in that The determining of the final risk value of the digital wallet based on the name of the digital wallet, the recharge behavior, the withdrawal behavior, the transfer receiving behavior, the login behavior, the number of times it is marked and the abnormal list screening result of the digital wallet includes: Determining a sensitivity score based on the name of the digital wallet; Determining an abnormal recharge score of the digital wallet based on the recharge behavior of the digital wallet account within a first preset time period; Determining an abnormal withdrawal score of the digital wallet based on the withdrawal behavior of the digital wallet account within a second preset time period; Determining an abnormal transfer score of the digital wallet based on the transfer receiving behavior of the digital wallet account within a third preset time period; Determining an abnormal login score of the digital wallet account based on the login behavior of the digital wallet account on an uncommon device; Determining a greylist screening score for the digital wallet account based on the number of times the digital wallet is marked; Determining an abnormal list screening score for the digital wallet account based on the abnormal list screening result; A final risk value is determined based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score.

3. The method according to claim 1, characterized in that The performing transaction risk control on the digital wallet account based on the risk level includes: If a transaction is detected in the digital wallet account, determining a transaction limit of the digital wallet account based on the risk level and the transaction type of the transaction; Determine whether the transaction meets the transaction limit, if so, allow the transaction, if not, block the transaction.

4. The method according to claim 2, characterized in that: The determining of the final risk value based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score includes: Determine the weights corresponding to the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score, and the abnormal list screening score respectively; wherein the sum of all weights is 1; The sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score are weighted and summed based on various weights to obtain the final risk value of the digital wallet account.

5. The method according to claim 4, characterized in that The process of determining each weight includes: Determine various weight adjustment parameters; For each of the weight adjustment parameters, determining the classification accuracy and weight sequence corresponding to the weight adjustment parameter based on the weight introspection learning method; the weight sequence includes the respective weights; The maximum classification accuracy is determined from the classification accuracy and weight sequences corresponding to the weight adjustment parameters of the respective weights, and the weight sequence corresponding to the maximum classification accuracy is used as the final weight sequence.

6. A digital wallet risk account identification system, characterized in that: The system comprises: A risk value determination unit, configured to determine a final risk value of the digital wallet account based on the name of the digital wallet account, the recharge behavior, the withdrawal behavior, the transfer receiving behavior, the login behavior, the number of times it is marked, and the abnormal list screening result of the digital wallet; a risk level determination unit, configured to determine a risk level of the digital wallet account based on the final risk value; A risk control unit performs transaction risk control on the digital wallet account based on the risk level.

7. The system according to claim 6, characterized in that The risk value determination unit includes: A sensitivity score module for determining a sensitivity score based on a name of a digital wallet; An abnormal recharge scoring module, used to determine an abnormal recharge score of the digital wallet based on the recharge behavior of the digital wallet account within a first preset time period; An abnormal withdrawal score module, used to determine the abnormal withdrawal score of the digital wallet based on the withdrawal behavior of the digital wallet account within a second preset time period; an abnormal transfer score module, configured to determine an abnormal transfer score of the digital wallet based on the transfer receiving behavior of the digital wallet account within a third preset time period; an abnormal login score module, used to determine an abnormal login score of the digital wallet account based on the login behavior of the digital wallet account on an uncommon device; A greylist screening score module, which determines a greylist screening score of the digital wallet account based on the number of times the digital wallet is marked; An abnormal list screening score module, used to determine the abnormal list screening score of the digital wallet account based on the abnormal list screening result; A risk value determination module is used to determine a final risk value based on the sensitivity score, the abnormal recharge score, the abnormal withdrawal score, the abnormal transfer score, the abnormal login score, the gray list screening score and the abnormal list screening score.

8. The system according to claim 6, characterized in that The risk control unit comprises: a transaction type determination module, configured to determine a transaction limit of the digital wallet account based on the risk level and the transaction type of the transaction when a transaction is detected in the digital wallet account; A determination module, used to determine whether the transaction meets the transaction limit; The transaction control module is used to allow the transaction when the transaction meets the transaction limit, and to block the transaction when the transaction does not meet the transaction limit.

9. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 5.

10. A computer storage medium, characterized in that: A computer program is stored thereon, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images