CA3161250A1 - Merchant tenant risk monitoring method, device, computer equipment and storage medium - Google Patents

Abstract

The present application relates to a merchant tenant risk monitoring method, and corresponding device, computer equipment and storage medium. The method comprises:
calculating to obtain risk assessment values before settlement according to similarity values between a to-be-settled merchant tenant and various risky merchant tenants, and obtaining platform record data formed by platform operations carried out by various target merchant tenants within a buffer period after settlement; performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond; employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond.

Description

MERCHANT TENANT RISK MONITORING METHOD, DEVICE, COMPUTER
EQUIPMENT AND STORAGE MEDIUM
BACKGROUND OF THE INVENTION
Technical Field [0001] The present application relates to the field of computer data processing technology, and more particularly to a merchant tenant risk monitoring method, and corresponding device, computer equipment and storage medium.
Description of Related Art

[0002] With the rapid development of the e-commerce field, more and more transactions are carried out online over the internet. At present, quality check of to-be-settled merchant tenants for applications to settle on platforms (also referred to as e-commerce platforms) appears particularly important, as this is closely related to the normal operation and risk control of the platforms.

[0003] For quite a long time, risk control at the level of merchant tenants is invariably highly dependent on human investigation and human check and analysis no matter whether it is in the banking industry, in the science and technology industry or in the financial industry.
Such a mode of over-dependence on manual collection of data and subjective judgment requires the consumption of lot of time and is not high in accuracy.
SUMMARY OF THE INVENTION

[0004] In view of the above technical problems, there is an urgent need to provide a merchant tenant risk monitoring method, and corresponding device, computer equipment and Date Recue/Date Received 2022-06-01 storage medium that are accurate and time-saving.

[0005] There is provided a merchant tenant risk monitoring method that comprises:

[0006] obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;

[0007] performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond;

[0008] employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and

[0009] calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

[0010] In one of the embodiments, the method further comprises:

[0011] obtaining basic information data of a to-be-settled merchant tenant applying for settlement on the platform;

[0012] performing feature extraction on the basic information data, and obtaining a feature vector before settlement of the to-be-settled merchant tenant;

[0013] performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with a feature vector to which each of plural risky merchant tenants corresponds, and obtaining similarity values between the to-be-settled merchant tenant and the various risky merchant tenants;

[0014] calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values between the to-be-settled merchant tenant and the various risky merchant tenants; and Date Recue/Date Received 2022-06-01

[0015] determining the to-be-settled merchant tenant as a target merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is not greater than a preset first risk assessment threshold.

[0016] In one of the embodiments, the method further comprises:

[0017] sending check prompt information to a terminal for prompting to further assess a risk of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the first risk assessment threshold and is not greater than a preset second risk assessment threshold; and

[0018] intercepting an application for settlement on the platform of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the second risk assessment threshold.

[0019] In one of the embodiments, the method is characterized in that:

[0020] the basic information data includes registration record information and external transfer information of the to-be-settled merchant tenant; and that

[0021] the platform record data includes a merchant business poi __________ tiait, transaction flow data, and order record data of the target merchant tenant within the buffer period after settlement.

[0022] In one of the embodiments, the various risky merchant tenants are subordinate to clusters corresponding thereto, and the step of calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values includes:

[0023] calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the following formula:
In4A5 AtoS .................. S MAn r=1 Rm C k

[0024]

[0025] where Rm c k expresses the risk assessment value before settlement of the to-be-settled merchant tenant, k expresses the number of clusters, Sm]Sw2 ...............
Smy, 1 expresses Date Recue/Date Received 2022-06-01 products between n i number of similarity values obtained after performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with feature vectors to which n i number of risky merchant tenants in the ith cluster correspond, and n i expresses the total number of risky merchant tenants subordinate to the ith cluster.

[0026] In one of the embodiments, the abnormality detection model is an isolation forest model, and the step of employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond includes:

[0027] inputting the feature matrix into the isolation forest model, outputting abnormality scores to which the various target merchant tenants correspond, numerically transforming the abnormality scores of the various target merchant tenants, and obtaining risk assessment values after settlement of the various target merchant tenants, wherein the risk assessment values after settlement are valuated in the range of being not smaller than 0 and not greater than 1.

[0028] In one of the embodiments, the step of calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond includes:

[0029] respectively weighting and thereafter summating the risk assessment values before and after settlement to which the various target merchant tenants correspond, and obtaining the comprehensive risk values to which the various target merchant tenants correspond.

[0030] There is provided a computer equipment that comprises a memory, a processor and a computer program stored on the memory and operable on the processor, and the following Date Recue/Date Received 2022-06-01 steps are realized when the processor executes the computer program:

[0031] obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;

[0032] performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond;

[0033] employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and

[0034] calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

[0035] There is provided a computer-readable storage medium storing a computer program thereon, and the following steps are realized when the computer program is executed by a processor:

[0036] obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;

[0037] performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond;

[0038] employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and

[0039] calculating comprehensive risk values to which the various target merchant tenants Date Recue/Date Received 2022-06-01 correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

[0040] In the aforementioned merchant tenant risk monitoring method, device, computer equipment and storage medium, by a series of such steps as performing feature extraction on the platform record data of the target merchant tenants in the buffer period after settlement, constructing a feature matrix, and detecting abnormality, risk assessment values after settlement are obtained, and it is made possible, in combination with risk assessment values before settlement, to determine comprehensive risk values of the target merchant tenants at the end of the buffer period after settlement, to timely monitor merchant tenant risks, and to obtain accurate monitoring results.
BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Fig. 1 is a view illustrating the application environment for the merchant tenant risk monitoring method in an embodiment;

[0042] Fig. 2 is a flowchart schematically illustrating the merchant tenant risk monitoring method in an embodiment;

[0043] Fig. 3 is a flowchart schematically illustrating the step of judging the size of a risk assessment value before settlement involved in an embodiment;

[0044] Fig. 4 is a block diagram illustrating the structure of the merchant tenant risk monitoring device in an embodiment;

[0045] Fig. 5 is a block diagram illustrating the structure of the risk assessment value calculating module before settlement involved in an embodiment; and Date Recue/Date Received 2022-06-01

[0046] Fig. 6 is a view illustrating the internal structure of the computer equipment in an embodiment.
DETAILED DESCRIPTION OF THE INVENTION

[0047] To make more lucid and clear the objectives, technical solutions and advantages of the present application, the present application is described in greater detail below with reference to accompanying drawings and embodiments. As should be understood, the specific embodiments described here are merely meant to explain the present application, rather than to restrict the present application.

[0048] The merchant tenant risk monitoring method provided by the present application is applicable to the application environment as shown in Fig. 1, in which server 101 can communicate with terminal 102 through network, for instance, after having calculated the compressive risk value of a target merchant tenant, server 101 can send the compressive risk value to terminal 102 for display by terminal 102, at this time, the checking personnel at the platform can learn of the risk situation of the target merchant tenant through the compressive risk value displayed by terminal 102. Server 101 can be embodied as an independent server or a server cluster consisting of a plurality of servers, and terminal 102 can be, but is not limited to be, any of various personal computers, notebook computers, smart mobile phones, panel computers, and portable wearable devices.

[0049] In one embodiment, as shown in Fig. 2, there is provided a merchant tenant risk monitoring method, and the method is explained with an example of its being applied to server 101 in Fig. 1, to comprise the following steps:

[0050] Step S201 - obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;

Date Recue/Date Received 2022-06-01

[0051] Step S202 - performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond;

[0052] Step S203 - employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and

[0053] Step S204 - calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

[0054] When a merchant tenant is not settled on the platform, this merchant tenant can be referred to as a to-be-settled merchant tenant, after an application of the merchant tenant to settle on the platform has been ratified, the merchant tenant can be referred to as a target merchant tenant. In the above steps, after target merchant tenants have settled on the platform, a buffer period after settlement should be undergone, under certain more specific circumstances, if after the buffer period after settlement ends, the comprehensive risk values of these target merchant tenants are within a reasonable range, then the target merchant tenants will be allowed to retain the settled status.

[0055] As should be noted, in step S201, in obtaining the two different types of data as the risk assessment values before settlement and the platform record data of specific target merchant tenants, it is possible to firstly obtain one type of data and then obtain another type of data, and it is also possible to simultaneously obtain the two different types of data.

[0056] The risk assessment values before settlement of the target merchant tenants can be determined according to the relevant data of to-be-settled merchant tenants when they are not settled. The specific obtaining mode will be described later in this paper.

Date Recue/Date Received 2022-06-01

[0057] The platform record data of the target merchant tenants at least includes data, such as order information or transaction flow information, etc., generated through commercial activities carried out by the target merchant tenants through the platform after settlement on the platform. However, under some specific circumstances, the platform record data can further include other data, and these data can be formed before settlement of to-be-settled merchant tenants on the platform, such as places of registration of enterprises to which the target merchant tenants correspond.

[0058] As regards the buffer period after settlement, it is a time period whose corresponding time duration can be set according to practical requirement, such as 10 days, half a month, or a full month, etc. The buffer period after settlement can be counted from the time point at which an application submitted by a to-be-settled merchant tenant to the platform for settlement on the platform is ratified, and can also be counted from a certain time point after settlement (for instance, from the time at which the first transaction occurs after the target merchant tenant has settled on the platform), to which no particular restriction is made in this context. The significance of setting up a buffer period after settlement lies in the facilitation to timely carry out risk surveillance after the target merchant tenant has settled on the platform, in the collection of relevant data corresponding to this stage, and in the facilitation to more accurately obtain the monitoring result (such as the comprehensive risk value) in conjunction with data of the to-be-settled merchant tenant before settlement on the platform.

[0059] As regards feature extraction, since it is needed to monitor the risks of various target merchant tenants, it is therefore required to process the platform record data of the various target merchant tenants before the monitoring, and the specific mode is the feature extraction. After feature extraction has been performed on the platform record data of each target merchant tenant, a feature vector after settlement corresponding to each target merchant tenant can be obtained. Feature extraction can be embodied as a currently Date Recue/Date Received 2022-06-01 available feature engineering algorithm, to which not much restriction is made in this context.

[0060] As regards the abnormality detection model, it is usually embodied as an unsupervised learning model, such as an isolation forest model, a KNN (K Nearest Neighbors) model, or an LOF (Local Outlier Factor) model, etc. The objective of abnormality processing rests mainly in differentiating data that is inconsistent with normal samples, to thereby determine any target merchant tenant who possesses abnormal behavior/property.
With respect to plural target merchant tenants that settle on a platform during a certain time period, the feature vector after settlement of each target merchant tenant is obtained to construct a feature matrix, and abnormality detection processing is performed on the feature matrix, whereby an abnormality score of each target merchant tenant can be obtained, and sizes of the abnormality scores reflect the levels of risks of behaviors of the target merchant tenants during the buffer period after settlement. Under some circumstances, the abnormality scores can be directly taken to serve as risk assessment values after settlement; under other circumstances, it is also possible to transform the abnormality scores to obtain risk assessment values after settlement possessing mapping relations to the abnormality scores.

[0061] The comprehensive risk values in step S204 are one piece of data that reflects comprehensive risks of target merchant tenants at the end of the buffer period after settlement. It is generally possible to respectively weight the risk assessment values before and after settlement and thereafter summate the two values, and then take the numerical values of such summation as comprehensive risk values. However, the use of other calculating modes to obtain the comprehensive risk values is not excluded, it is required at this time to provide the risk assessment values before settlement and the comprehensive risk values with correlation, and to simultaneously provide the risk assessment values after settlement and the comprehensive risk values also with correlation, such correlations can be positive correlations, moreover, such correlations Date Recue/Date Received 2022-06-01 can be proportional correlations.

[0062] In the aforementioned merchant tenant risk monitoring method, by a series of such steps as performing feature extraction on the platform record data of the target merchant tenants within the buffer period after settlement, constructing a feature matrix, and detecting abnormality, risk assessment values after settlement are obtained, and it is made possible, in combination with risk assessment values before settlement, to determine comprehensive risk values of the target merchant tenants at the end of the buffer period after settlement. Data generated within the time interval of the buffer period after settlement is utilized to perform risk surveillance, whereby risk results of merchant tenants after settlement on the platform can be timely obtained, risk assessment values after settlement and risk assessment values before settlement are processed, risks reflected by properties or behaviors of merchant tenants before settlement on the platform are not only considered, but risks reflected by properties or behaviors of merchant tenants after settlement on the platform are also considered, the comprehensive risk values obtained thereby can reflect comprehensive risk levels of target merchant tenants from an overall perspective, thus facilitating to enhance accuracy of risk monitoring results of merchant tenants. As should be noted, the risk monitoring results, or monitoring results, as mentioned in this paper mainly stand for data descriptive of risk levels of merchant tenants (including target merchant tenants) directly or indirectly obtained by the merchant tenant risk monitoring method, the directly obtained data can be compressive risk values, risk assessment values before settlement, or assessment values after settlement, and the indirectly obtained data can be obtained by mapping processing of the directly obtained data, for instance, the directly obtained data can be mapping-processed to obtain corresponding risk levels, and the risk levels can as well be regarded as monitoring results at this time.

[0063] In some embodiments, as shown in Fig. 3, the merchant tenant risk monitoring method further comprises:

Date Recue/Date Received 2022-06-01

[0064] Step S301 - obtaining basic information data of a to-be-settled merchant tenant applying for settlement on the platform;

[0065] Step S302 - performing feature extraction on the basic information data, and obtaining a feature vector before settlement of the to-be-settled merchant tenant;

[0066] Step S303 - performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with a feature vector to which each of plural risky merchant tenants corresponds, and obtaining similarity values between the to-be-settled merchant tenant and the various risky merchant tenants;

[0067] Step S304 - calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values between the to-be-settled merchant tenant and the various risky merchant tenants; and

[0068] Step S305 - determining the to-be-settled merchant tenant as a target merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is not greater than a preset first risk assessment threshold.

[0069] As regards server 101, there will be the circumstance in which applications for settlement on the platform will be incessantly received. In this case, when the risk assessment value before settlement is not greater than a first risk assessment threshold, the to-be-settled merchant tenant to which the risk assessment value before settlement corresponds can be determined as a target merchant tenant by the execution of steps S301-S305, and risk evaluation is performed once again at the end of the buffer period after settlement of the target merchant tenant by the execution of steps S201-S204, to obtain the comprehensive risk value. It is thusly not only ensured that the merchant tenant being settled on the platform has relatively low risk, but also made possible to carry out further risk surveillance to the merchant tenant being settled on the platform.

[0070] The feature extraction in step S302 can be carried out by a currently available feature engineering algorithm mentioned before, to which not much restriction is made in this context.

Date Recue/Date Received 2022-06-01

[0071] The risky merchant tenants in step S303 mainly indicate suspect or abnormal merchant tenant samples currently in grasp, as regards the risky merchant tenants, it is also possible to employ the mode of feature extraction to obtain feature vectors corresponding to the risky merchant tenants. When the number of risky merchant tenants is relatively large, the risky merchant tenants can be marked, risky merchant tenants with the same mark are considered to be one cluster, and such marks can be designated with correspondences with such behaviors/properties as operation anomaly, pornography/gambling/drugging, click farming, and cashing, etc. Performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with a feature vector to which each of plural risky merchant tenants corresponds can mainly include the following two modes:

[0072] (1) Similarity calculation is performed on the feature vector before settlement with feature vectors of risky merchant tenants in one cluster, to obtain similarity values corresponding to this cluster; in accordance with this mode, similarity values of the feature vector before settlement with respect to plural clusters can be obtained, generally speaking, how many clusters there are, that many similarity values can be obtained; on the basis thereof, the plural similarity values are averaged to finally obtain an average value, and this average value can be regarded as the similarity value between the to-be-settled merchant tenant and the various risky merchant tenants.

[0073] (2) Calculation is performed not in accordance with the cluster mode, whereas uniform similarity calculation is directly performed on the feature vector before settlement and the feature vectors of all risky merchant tenants, and similarity values between the to-be-settled merchant tenant and the various risky merchant tenants are directly obtained.

[0074] Determining the to-be-settled merchant tenant as a target merchant tenant in step S305 can be either automatically executed by server 101, or triggered by server 101 according to an instruction input by the platform checking personnel and sent by terminal 102. It is specifically possible to add the settled merchant tenant from a list of settlement check to Date Recue/Date Received 2022-06-01 a list of check approval. Alternatively, the mark of application for settlement of the to-be-settled merchant tenant is changed to a mark corresponding to the platform settled.

[0075] In some embodiments, as shown in Fig. 3, the merchant tenant risk monitoring method further comprises:

[0076] Step S306 - sending check prompt information to terminal 102 for prompting to further assess a risk of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the first risk assessment threshold and is not greater than a preset second risk assessment threshold;
and

[0077] Step S307 - intercepting an application for settlement on the platform of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the second risk assessment threshold.

[0078] The first risk assessment threshold and the second risk assessment threshold in Fig. 3 can be designed according to practical requirement, and no particular restriction is made thereto. The check prompt information functions mainly to notify the platform checking personnel to assess the risk of the to-be-settled merchant tenant. Moreover, terminal 102 can detect a feedback instruction of the platform checking personnel, and sends the feedback instruction to server 101; if the feedback instruction is for instructing server 101 to determine the to-be-settled merchant tenant as a target merchant tenant, server 101 executes the processing steps to determine the to-be-settled merchant tenant as a target merchant tenant; if the feedback instruction is for instructing server 101 to intercept an application for settlement on the platform of the to-be-settled merchant tenant, server 101 executes the interception processing step. Thusly, it is facilitated to make more accurate judgment on the risk of the to-be-settled merchant tenant.

[0079] When the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the second risk assessment threshold, the application for settlement on the platform of the to-be-settled merchant tenant is intercepted, indicating that the risk of the Date Recue/Date Received 2022-06-01 to-be-settled merchant tenant is relatively high, at this time it is not required for the platform checking personnel to check, as it is possible for server 101 to make the intercepting process on its own initiative. The intercepting process is a process to not allow settlement of the application for settlement on the platform, whereby the to-be-settled merchant tenant cannot be settled on the platform at least within a period of time.
With respect to such intercepting process, server 101 can generate first interception prompt information and send the same terminal 102, so as to let the platform checking personnel learn of the event.

[0080] In some embodiments, the basic information data includes registration record information and external transfer information of the to-be-settled merchant tenant, and the platform record data includes a merchant business poi ______________________________ ti aft, transaction flow data, and order record data of the target merchant tenant within the buffer period after settlement.

[0081] Generally speaking, since the dimensions of the obtainable data are limited before the to-be-settled merchant tenant is settled on the platform, the basic information data at this time can include registration record information, region information, and legal person qualification information of the to-be-settled merchant tenant. Further, the registration record information includes, but is not limited to, registered mailbox information, registered telephone information, and registration time information. The basic information data can further include external transfer information, and the external transfer information includes, but is not limited to, transfer frequency information, actual controller transfer information, and transfer bank information. However, the platform record data is different, as it is mainly derived from commercial behavior data of target merchant tenants recorded by the server to which the platform corresponds, such as customer transactions, transfer information, and deposit and withdrawal information, etc.
As should be noted, in addition to the data generated or formed within the buffer period after settlement, the platform record data can further include data generated or formed before settlement on the platform. Under some specific circumstances, the merchant Date Recue/Date Received 2022-06-01 business poi ______________________________________________________________ Li aft includes, but is not limited to, customer type (corporate-oriented type, privately oriented type, etc.) data, registered capital data, data reflecting whether the account of a target merchant tenant exists and whether the registered mobile phone region and registered account region do not conform to each other, and data reflecting whether the registered region of the target merchant tenant originates from a highly risky region.
The transaction flow data can be transaction flow slice information, which mainly involves money flow and flow time slice information of the target merchant tenant within the buffer period after settlement, such as balance information of associated accounts at each time point within half a month after settlement on the platform; the transaction flow slice information can also include information representing flowing in and out of great amount of capital, abrupt change in account balance, unduly much orders in abnormal transaction times, unduly much transfer of orders with similar amounts of money, unduly much amount of transfer across the border or across provinces, or frequent cash withdrawal/transfer of corporate accounts, etc. The order record data mainly includes online/offline order record information of the target merchant tenant after settlement on the platform, focusing principally on portrait information of transactions, including, but not limited to, transaction frequency information, machine fingerprint information, order amount of money information, and buyer qualification record information.

[0082] In some embodiments, when the number of clusters to which risky merchant tenants correspond is k, Ci [m,õm] can be employed to represent a set of all risky merchant tenants in the ith cluster, where 1 i<k , mo represents the first risky merchant tenant in the ith cluster, by the same token, m,õ 1 represents the n ith risky merchant tenant in the ith cluster, and n i represents the number of risky merchant tenants in this cluster. At this time, with respect to a certain to-be-settled merchant tenant (represented by M), feature extraction can be performed on its basic information data, to obtain d dimension of a feature vector before settlement (represented by f, ), Date Recue/Date Received 2022-06-01 fM = (t,õ/õt ); with respect to a certain risky merchant tenant (represented by min ), feature extraction can be performed on the basic information data of this risky merchant tenant, to obtain d dimension of a feature vector to which this risky merchant tenant corresponds, represented by fm ,,j, fm,,=

[0083] To facilitate further understanding, explanation is made with a circumstance in which the number of clusters is 2, in one of which cluster the number of risky merchant tenants is 2, and in another one of which cluster the number of risky merchant tenants is 3. At this time, k is 2, with respect to the first cluster, i is valuated as 1, n = n 1 =
2, the set of risky merchant tenants in the first cluster is m1,2] , where rn11 represents the first risky merchant tenant in the first cluster, and mu represents the second risky merchant tenant in the first cluster. With respect to the second cluster, i is valued as 2, n i = n 2= 3 , the set of risky merchant tenants in the second cluster is C2[11/2,/, m2,2, m2,3] m21 represents the first risky merchant tenant in the second cluster, M2,2 represents the second risky merchant tenant in the second cluster, and m23 represents the third risky merchant tenant in the second cluster. Suppose d=6, then as regards rn11 (the first risky merchant tenant in the first cluster), its corresponding feature vector is , = (tinny mll; the feature vectors to which other risky merchant tenants correspond can be represented in accordance with the similar principle, to which no repetition is made in this context.

[0084] Performing similarity matching calculation on the feature vector before settlement of a certain to-be-settled merchant tenant (represented by M) with a feature vector of a certain risky merchant tenant (represented by in,n ) can be to calculate their cosine similarity value. The calculation formula of the cosine similarity value can be inferred from the Date Recue/Date Received 2022-06-01 following formula (1).

[0085] Consine similarity(a = b) ¨ a b (1) MaMMbM

[0086] In formula (1), a and b respectively represent two pieces of data of the cosine similarity value to be calculated, specifically, they can respectively represent the feature vector before settlement and the feature vector of the risky merchant tenant.
Consine similarity(a = b) represents the cosine similarity value of the feature vector before settlement and the feature vector of the risky merchant tenant.

[0087] Thus, with respect to a certain cluster Ci[mi,õ , the similarity values of the feature vector fm before settlement of a certain to-be-settled merchant tenant and the feature vectors of risky merchant tenants in this cluster can be represented as (Smo õSmo, , where Smo represents the similarity value (such as a cosine similarity value) of fm with the feature vector of the first risky merchant tenant in this cluster obtained after similarity matching calculation, by the same token, Smo represents the similarity value of fm with the feature vector of the n ith risky merchant tenant in this cluster obtained after similarity matching calculation.

[0088] In some embodiments, the various risky merchant tenants are subordinate to clusters corresponding thereto, and step S304 includes calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to formula (2):
In4r5At]SAu2 ................. S Mn r

[0089] Rm c k = __ 1=1 (2) Date Recue/Date Received 2022-06-01

[0090] where Rm õ expresses the risk assessment value before settlement of the to-be-settled merchant tenant, it is also a comprehensive similarity value of the feature vector before settlement of the to-be-settled merchant tenant with the feature vectors to which the risky merchant tenants of k number of clusters, k expresses the number of clusters, ..........................................................................
Sm,õ 1 expresses products between n i number of similarity values obtained after performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with feature vectors to which n i number of risky merchant tenants in the ith cluster correspond, and n i expresses the total number of risky merchant tenants subordinate to the ith cluster.

[0091] To facilitate comprehension, an example is provided here. In this example, k=2. Suppose when i=1, n i=n 1= 2; suppose when i=2, n i=n 2= 3 .

[0092] Then there is R
m C k All] All' +mziSm22Sm2,3

[0093] Seen as such, JS 1S. S _________________________________________ 1 min represents a geometrical average similarity value of ................................ õ
fm with plural similarity values of feature vectors of the risky merchant tenants in the cluster, moreover, Rm k represents a compressive similarity value obtained after employing an arithmetic averaging calculation on plural geometrical average similarity values.

[0094] In some embodiments, Rm k can be regarded as a risk assessment value before settlement, its valuation range lies between 0 and 1, the higher this value is, the higher will be the risk level of the to-be-settled merchant tenant.

Date Recue/Date Received 2022-06-01

[0095] In some embodiments, the abnormality detection model is an isolation forest model, and step S203 includes: inputting the feature matrix into the isolation forest model, outputting abnormality scores to which the various target merchant tenants correspond, numerically transforming the abnormality scores of the various target merchant tenants, and obtaining risk assessment values after settlement of the various target merchant tenants, wherein the risk assessment values after settlement are valuated in the range of being not smaller than 0 and not greater than 1.

[0096] Specifically, a currently available isolation forest model can be used, for example, an Isolation Forest function in python is used to serve as the function to which the isolation forest model corresponds, at which time the feature matrix is substituted in the function, and abnormality scores in the valuation range of -1 to 1 will be obtained.

[0097] Numerically transforming the abnormality scores of the target merchant tenants aims to correspond the valuation range of the risk assessment value after settlement to the valuation range of the risk assessment value before settlement, and the two valuation ranges both lie within 0 and 1, so as to facilitate subsequent calculation of the comprehensive risk value.

[0098] Suppose that n number of merchant tenants in a group of to-be-settled merchant tenants is ratified for settlement, that is to say, there are n target merchant tenants now, and these target merchant tenants can be represented by M1, M,õ Mn , in which Mi represents the first target merchant tenant, so on and so forth, and Mn represents the nth target merchant tenant. Feature extraction is performed on the platform record data of these target merchant tenants respectively, whereby feature vectors after settlement of n number of target merchant tenants can be obtained, suppose that the dimension of each feature vector after settlement is p, a matrix is constructed out of the feature vectors after settlement of the n number of target merchant tenants, and a feature matrix can be obtained (represented by X), Date Recue/Date Received 2022-06-01 X11 = = = Xlp X= = = .
= = = x

[0099] X1 nP /

[0100] Element xnp of X represents the pth feature value of the nth target merchant tenant (namely Ma).

[0101] X is input in the abnormality detection model, and abnormality scores of n number of target merchant tenants can be output, with respect to a certain target merchant tenant (represented by M , 1<j<n ), its abnormality score is t, , since -1 ti , so numerical transformation can be performed according to formula (3).
V-(1 -1)

[0102] P ¨ (2)

[0103] where Pi represents a risk assessment value after settlement corresponding to M1, and its valuation range lies between 0 to 1.

[0104] In some embodiments, step S204 includes: respectively weighting and thereafter summating the risk assessment values before and after settlement to which the various target merchant tenants correspond, and obtaining the comprehensive risk values to which the various target merchant tenants correspond. More specifically, under some circumstances, the comprehensive risk value to which M corresponds is represented by RiskScore MJ, which can be calculated and obtained according to formula (4).

[0105] RiskScore MJ =a = Rm k+ (1-a )=P (4) ¨

[0106] where a represents a weight to which Rm, k corresponds, at this time, Rm, k Date Recue/Date Received 2022-06-01 represents the risk assessment value before settlement to which Mi corresponds, and (1- a) represents a weight to which Pj corresponds.

[0107] In some embodiments, with respect to a certain target merchant tenant, when its corresponding comprehensive risk value is greater than a preset comprehensive risk threshold, server 101 sends second interception information (the second interception information is used to prompt that the risk of the target merchant tenant is relatively high) to terminal 102, and performs an intercepting process of the target merchant tenant after settlement on the platform, for instance, to intercept the corresponding online shop of the target merchant tenant or put the corresponding commodities of the target merchant tenant off the shelf. If the comprehensive risk value is not greater than the preset comprehensive risk threshold, it is possible to update the platform record data again after a designated time duration, and to calculate again the comprehensive risk value of the target merchant tenant, when the updated comprehensive risk value is greater than the preset comprehensive risk threshold, server 101 performs an intercepting process on the target merchant tenant after settlement on the platform.

[0108] As should be understood, although the various steps in the flowcharts of Figs. 2 and 3 are sequentially displayed as indicated by arrows, these steps are not necessarily executed in the sequences indicated by arrows. Unless otherwise explicitly noted in this paper, execution of these steps is not restricted by any sequence, as these steps can also be executed in other sequences (than those indicated in the drawings). Moreover, at least partial steps in the flowcharts of Figs. 2 and 3 may include plural sub-steps or multi-phases, these sub-steps or phases are not necessarily completed at the same timing, but can be executed at different timings, and these sub-steps or phases are also not necessarily sequentially performed, but can be performed in turns or alternately with other steps or with at least some of sub-steps or phases of other steps.

Date Recue/Date Received 2022-06-01

[0109] In one embodiment, as shown in Fig. 4, there is provided a merchant tenant risk monitoring device 400 that comprises:

[0110] a data obtaining module 401, for obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;

[0111] a feature extracting module 402, for performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond;

[0112] an abnormality detecting module 403, for employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and

[0113] a risk calculating module 404, for calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

[0114] In one embodiment, as shown in Fig. 5, the merchant tenant risk monitoring device 400 further comprises:

[0115] a basic information data obtaining module 501, for obtaining basic information data of a to-be-settled merchant tenant applying for settlement on the platform;

[0116] a feature vector before settlement obtaining module 502, for performing feature extraction on the basic information data, and obtaining a feature vector before settlement of the to-be-settled merchant tenant;

[0117] a similarity matching and calculating module 503, for performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with a feature vector to which each of plural risky merchant tenants corresponds, and obtaining similarity values between the to-be-settled merchant tenant and the various Date Recue/Date Received 2022-06-01 risky merchant tenants;

[0118] a risk assessment value before settlement calculating module 504, for calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values between the to-be-settled merchant tenant and the various risky merchant tenants; and

[0119] a target merchant tenant determining module 505, for determining the to-be-settled merchant tenant as a target merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is not greater than a preset first risk assessment threshold.

[0120] In one embodiment, as shown in Fig. 5, the merchant tenant risk monitoring device 400 further comprises:

[0121] a check prompt information sending module 506, for sending check prompt information to a terminal for prompting to further assess a risk of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the first risk assessment threshold and is not greater than a preset second risk assessment threshold; and

[0122] an interception processing module 507, for intercepting an application for settlement on the platform of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the second risk assessment threshold.

[0123] In one embodiment, the risk assessment value before settlement calculating module calculates to obtain the risk assessment value before settlement of the to-be-settled merchant tenant according to formula (2).

[0124] In one embodiment, the abnormality detection model is an isolation forest model, and the abnormality detecting module 403 inputs the feature matrix in the isolation forest model, outputs abnormality scores to which the various target merchant tenants correspond, Date Recue/Date Received 2022-06-01 numerically transforms the abnormality scores of the various target merchant tenants, and obtains risk assessment values after settlement of the various target merchant tenants, wherein the risk assessment values after settlement are valuated in the range of being not smaller than 0 and not greater than 1.

[0125] In one embodiment, the risk calculating module 404 respectively weights and thereafter summates the risk assessment values before and after settlement to which the various target merchant tenants correspond, and obtains the comprehensive risk values to which the various target merchant tenants correspond.

[0126] Specific definitions relevant to the merchant tenant risk monitoring device may be inferred from the aforementioned definitions to the merchant tenant risk monitoring method, while no repetition is made in this context. The various modules in the aforementioned merchant tenant risk monitoring device can be wholly or partly realized via software, hardware, and a combination of software with hardware. The various modules can be embedded in the form of hardware in a processor in a computer equipment or independent of any computer equipment, and can also be stored in the form of software in a memory in a computer equipment, so as to facilitate the processor to invoke and perform operations corresponding to the aforementioned various modules.

[0127] In one embodiment, a computer equipment is provided, the computer equipment can be a server, and its internal structure can be as shown in Fig. 6. The computer equipment comprises a processor, a memory, and a network interface connected to each other via a system bus. The processor of the computer equipment is employed to provide computing and controlling capabilities. The memory of the computer equipment includes a nonvolatile storage medium and an internal memory. The nonvolatile storage medium stores therein an operating system, and a computer program. The internal memory provides environment for the running of the operating system and the computer program in the nonvolatile storage medium. The network interface of the computer equipment is Date Recue/Date Received 2022-06-01 employed to connect to an external terminal via network for communication. The computer program realizes a merchant tenant monitoring method when it is executed by a processor.

[0128] As understandable to persons skilled in the art, the structure illustrated in Fig. 6 is merely a block diagram of partial structure relevant to the solution of the present application, and does not constitute any restriction to the computer equipment on which the solution of the present application is applied, as the specific computer equipment may comprise component parts that are more than or less than those illustrated in Fig. 6, or may combine certain component parts, or may have different layout of component parts.

[0129] In one embodiment, there is provided a computer equipment that comprises a memory, a processor and a computer program stored on the memory and operable on the processor, and the following steps are realized when the processor executes the computer program:
obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;
performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond; employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

[0130] In one embodiment, when the processor executes the computer program, the following steps are further realized: obtaining basic information data of a to-be-settled merchant Date Recue/Date Received 2022-06-01 tenant applying for settlement on the platform; performing feature extraction on the basic information data, and obtaining a feature vector before settlement of the to-be-settled merchant tenant; performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with a feature vector to which each of plural risky merchant tenants corresponds, and obtaining similarity values between the to-be-settled merchant tenant and the various risky merchant tenants;
calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values between the to-be-settled merchant tenant and the various risky merchant tenants; and determining the to-be-settled merchant tenant as a target merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is not greater than a preset first risk assessment threshold.

[0131] In one embodiment, when the processor executes the computer program, the following steps are further realized: sending check prompt information to a terminal for prompting to further assess a risk of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the first risk assessment threshold and is not greater than a preset second risk assessment threshold;
and intercepting an application for settlement on the platform of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the second risk assessment threshold.

[0132] In one embodiment, the various risky merchant tenants are subordinate to clusters corresponding thereto, and the step of calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values is realized when the processor executes the computer program includes:
calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the formula (2).

[0133] In one embodiment, the abnormality detection model is an isolation forest model, and the Date Recue/Date Received 2022-06-01 step of employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond is realized when the processor executes the computer program includes: inputting the feature matrix into the isolation forest model, outputting abnormality scores to which the various target merchant tenants correspond, numerically transforming the abnormality scores of the various target merchant tenants, and obtaining risk assessment values after settlement of the various target merchant tenants, wherein the risk assessment values after settlement are valuated in the range of being not smaller than 0 and not greater than 1.

[0134] In one embodiment, the step of calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond is realized when the processor executes the computer program includes: respectively weighting and thereafter summating the risk assessment values before and after settlement to which the various target merchant tenants correspond, and obtaining the comprehensive risk values to which the various target merchant tenants correspond.

[0135] In one embodiment, there is provided a computer-readable storage medium storing thereon a computer program, and the following steps are realized when the computer program is executed by a processor: obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement; performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond; employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target Date Recue/Date Received 2022-06-01 merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

[0136] In one embodiment, when the computer program is executed by a processor, the following steps are further realized: obtaining basic information data of a to-be-settled merchant tenant applying for settlement on the platform; performing feature extraction on the basic information data, and obtaining a feature vector before settlement of the to-be-settled merchant tenant; performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with a feature vector to which each of plural risky merchant tenants corresponds, and obtaining similarity values between the to-be-settled merchant tenant and the various risky merchant tenants;
calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values between the to-be-settled merchant tenant and the various risky merchant tenants; and determining the to-be-settled merchant tenant as a target merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is not greater than a preset first risk assessment threshold.

[0137] In one embodiment, when the computer program is executed by a processor, the following steps are further realized: sending check prompt information to a terminal for prompting to further assess a risk of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the first risk assessment threshold and is not greater than a preset second risk assessment threshold;
and intercepting an application for settlement on the platform of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the second risk assessment threshold.

Date Recue/Date Received 2022-06-01

[0138] In one embodiment, the various risky merchant tenants are subordinate to clusters corresponding thereto, and the step of calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values is realized when the computer program is executed by a processor includes calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the formula (2).

[0139] In one embodiment, the abnormality detection model is an isolation forest model, and the step of employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond is realized when the computer program is executed by a processor includes: inputting the feature matrix into the isolation forest model, outputting abnormality scores to which the various target merchant tenants correspond, numerically transforming the abnormality scores of the various target merchant tenants, and obtaining risk assessment values after settlement of the various target merchant tenants, wherein the risk assessment values after settlement are valuated in the range of being not smaller than 0 and not greater than 1.

[0140] In one embodiment, the step of calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond is realized when the computer program is executed by a processor includes: respectively weighting and thereafter summating the risk assessment values before and after settlement to which the various target merchant tenants correspond, and obtaining the comprehensive risk values to which the various target merchant tenants correspond.

[0141] As comprehensible to persons ordinarily skilled in the art, the entire or partial flows in the methods according to the aforementioned embodiments can be completed via a Date Recue/Date Received 2022-06-01 computer program instructing relevant hardware, the computer program can be stored in a nonvolatile computer-readable storage medium, and the computer program can include the flows as embodied in the aforementioned various methods when executed. Any reference to the memory, storage, database or other media used in the various embodiments provided by the present application can all include nonvolatile and/or volatile memory/memories. The nonvolatile memory can include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable and programmable ROM (EEPROM) or a flash memory. The volatile memory can include a random access memory (RAM) or an external cache memory. To serve as explanation rather than restriction, the RAM is obtainable in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM
(SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct Rambus dynamic RAM (DRDRAM), and Rambus dynamic RAM
(RDRAM), etc.

[0142] Technical features of the aforementioned embodiments are randomly combinable, while all possible combinations of the technical features in the aforementioned embodiments are not exhausted for the sake of brevity, but all these should be considered to fall within the scope recorded in the Description as long as such combinations of the technical features are not mutually contradictory.

[0143] The foregoing embodiments are merely directed to several modes of execution of the present application, and their descriptions are relatively specific and detailed, but they should not be hence misunderstood as restrictions to the inventive patent scope. As should be pointed out, persons with ordinary skill in the art may further make various modifications and improvements without departing from the conception of the present application, and all these should pertain to the protection scope of the present application.
Accordingly, the patent protection scope of the present application shall be based on the Date Recue/Date Received 2022-06-01 attached Claims.

Date Recue/Date Received 2022-06-01

Claims (10)

What is claimed is:

1. A merchant tenant risk monitoring method, characterized in that the method comprises:
obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;
performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond;
employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

2. The method according to Claim 1, characterized in that the method further comprises:
obtaining basic information data of a to-be-settled merchant tenant applying for settlement on the platform;
performing feature extraction on the basic information data, and obtaining a feature vector before settlement of the to-be-settled merchant tenant;
performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with a feature vector to which each of plural risky merchant tenants corresponds, and obtaining similarity values between the to-be-settled merchant tenant and the various risky merchant tenants;
calculating to obtain a risk assessment value before settlement of the to-be-settled merchant Date Recue/Date Received 2022-06-01 tenant according to the similarity values between the to-be-settled merchant tenant and the various risky merchant tenants; and determining the to-be-settled merchant tenant as a target merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is not greater than a preset first risk assessment threshold.

3. The method according to Claim 2, characterized in that the method further comprises:
sending check prompt information to a terminal for prompting to further assess a risk of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the first risk assessment threshold and is not greater than a preset second risk assessment threshold; and intercepting an application for settlement on the platform of the to-be-settled merchant tenant when the risk assessment value before settlement of the to-be-settled merchant tenant is greater than the second risk assessment threshold.

4. The method according to Claim 2, characterized in that:
the basic information data includes registration record information and external transfer information of the to-be-settled merchant tenant; and that the platform record data includes a merchant business poi _________________ tiait, transaction flow data, and order record data of the target merchant tenant within the buffer period after settlement.

5. The method according to Claim 2, characterized in that the various risky merchant tenants are subordinate to clusters corresponding thereto, and that the step of calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the similarity values includes:
calculating to obtain a risk assessment value before settlement of the to-be-settled merchant tenant according to the following formula:

Date Recue/Date Received 2022-06-01 where R, õ expresses the risk assessment value before settlement of the to-be-settled merchant tenant, k expresses the number of clusters, SmoS" ................
SAtin expresses products between n i number of similarity values obtained after performing similarity matching calculation on the feature vector before settlement of the to-be-settled merchant tenant with feature vectors to which n i number of risky merchant tenants in the ith cluster correspond, and n_i expresses the total number of risky merchant tenants subordinate to the ith cluster.

6. The method according to Claim 5, characterized in that the abnormality detection model is an isolation forest model, and that the step of employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond includes:
inputting the feature matrix into the isolation forest model, outputting abnormality scores to which the various target merchant tenants correspond, numerically transforming the abnormality scores of the various target merchant tenants, and obtaining risk assessment values after settlement of the various target merchant tenants, wherein the risk assessment values after settlement are valuated in the range of being not smaller than 0 and not greater than 1.

7. The method according to anyone of Claims 1 to 6, characterized in that the step of calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond includes:
respectively weighting and thereafter summating the risk assessment values before and after settlement to which the various target merchant tenants correspond, and obtaining the comprehensive risk values to which the various target merchant tenants correspond.

8. A merchant tenant risk monitoring device, characterized in that the device comprises:
a data obtaining module, for obtaining risk assessment values before settlement to which plural target merchant tenants respectively correspond, and platform record data formed by platform operations carried out by the various target merchant tenants within a buffer period after settlement;
a feature extracting module, for performing feature extraction on the platform record data of the various target merchant tenants respectively, and obtaining feature vectors after settlement to which the various target merchant tenants correspond;
an abnormality detecting module, for employing an abnormality detection model to perform abnormality detection processing on a feature matrix constituted by the feature vectors after settlement to which the plural target merchant tenants correspond, and obtaining risk assessment values after settlement to which the various target merchant tenants correspond; and a risk calculating module, for calculating comprehensive risk values to which the various target merchant tenants correspond according to the risk assessment values before and after settlement to which the various target merchant tenants correspond.

9. A computer equipment, comprising a memory, a processor and a computer program stored on the memory and operable on the processor, characterized in that steps of the method according to anyone of Claims 1 to 7 are realized when the processor executes the computer program.

10. A computer-readable storage medium, storing a computer program thereon, characterized in that steps of the method according to anyone of Claims 1 to 7 are realized when the computer program is executed by a processor.