Payment detection processing method and apparatus

Abstract

Embodiments of the present specification provide a payment detection processing method and device, wherein the payment detection processing method comprises: obtaining a protocol payment channel opened by an associated user to a payment user; performing interaction type detection based on first interaction data of the payment user and the associated user; if the detection fails, performing interaction attenuation calculation based on second interaction data of the payment user and the associated user; if the calculated interaction attenuation value is less than an attenuation threshold, calling a relationship prediction model to predict the relationship between the payment user and the associated user; and if it is determined that the relationship is removed based on the prediction result, performing marking processing on the protocol payment channel.

Description

TECHNICAL FIELD

[0001] The present document relates to the technical field of data processing, and particularly relates to a payment detection processing method and device. BACKGROUND

[0002] With the continuous development and promotion of the Internet, the application range of online payment based on the Internet is also more and more wide, and has gradually covered most users. In this case, various forms of online payment have appeared, such as mutual payment between users, which makes the online payment of users more convenient. However, at the same time, complex user relationships may cause risks in mutual payment between users. Therefore, how to improve the payment security in the online payment process and reduce the possibility of user asset loss has become a new test for payment platforms in providing payment services for users. SUMMARY

[0003] One or more embodiments of the present specification provide a payment detection processing method, comprising: obtaining a protocol payment channel opened by an associated user to a payment user. Performing interaction type detection based on first interaction data of the payment user and the associated user. If the detection fails, performing interaction attenuation calculation based on second interaction data of the payment user and the associated user. If the obtained interaction attenuation value is less than the attenuation threshold, calling a relationship prediction model to predict the relationship between the payment user and the associated user. If it is determined that the relationship is removed based on the prediction result, marking the protocol payment channel.

[0004] One or more embodiments of the present specification provide a second payment detection processing method applied to a user terminal of a payment user, the method comprising: submitting a payment access request to a server based on a payment operation of the payment user. Receiving and displaying a payment channel set containing a protocol payment channel issued by the server. The protocol payment channel is marked after interaction type detection, interaction attenuation calculation and relationship prediction between the payment user and an associated user who opens the protocol payment channel. If a payment instruction for the protocol payment channel is detected, a payment processing request is sent to the server to perform payment confirmation of the protocol payment channel for the associated user.

[0005] This specification provides one or more embodiments of a third payment detection processing method applied to a server. The method includes: obtaining the protocol payment channels opened by an associated user for the paying user based on a payment access request submitted by the paying user's user terminal; performing relationship detection processing between the paying user and the associated user based on interaction data between the paying user and the associated user; the relationship detection processing including interaction type detection, interaction decay calculation, and relationship termination prediction; if the relationship detection fails, marking the protocol payment channels and sending a set of payment channels carrying the protocol payment channels to the user terminal; and performing payment confirmation for the associated user based on the payment processing request sent by the user terminal for the protocol payment channels.

[0006] This specification provides one or more embodiments of a payment detection processing apparatus, comprising: a protocol payment channel acquisition module configured to acquire protocol payment channels opened by an associated user for a payment user; an interaction type detection module configured to perform interaction type detection based on first interaction data between the payment user and the associated user; an interaction attenuation calculation module configured to perform interaction attenuation calculation based on second interaction data between the payment user and the associated user if the detection fails; a relationship termination prediction module configured to invoke a relationship prediction model to predict the termination of the relationship between the payment user and the associated user if the calculated interaction attenuation value is less than an attenuation threshold; and a marking processing module configured to mark the protocol payment channels if the relationship is determined to be terminated based on the prediction result.

[0007] This specification provides one or more embodiments of a second payment detection and processing device, operating on a user terminal of a payment user. The device includes: a payment access request submission module configured to submit a payment access request to a server based on the payment user's payment operation; a payment channel display module configured to receive and display a set of payment channels containing protocol payment channels issued by the server; the protocol payment channels are marked after interaction type detection, interaction attenuation calculation, and relationship dissolution prediction for the payment user and associated users who have activated the protocol payment channels; and a payment processing request sending module configured to send a payment processing request to the server if a payment instruction for the protocol payment channel is detected, to confirm payment for the associated user through the protocol payment channel.

[0008] One or more embodiments of the present specification provide a third payment detection processing apparatus running on a server, the apparatus comprising: a protocol payment channel obtaining module configured to obtain a protocol payment channel opened by an associated user to a payment user according to a payment access request submitted by a user terminal of the payment user; a relationship detection processing module configured to perform relationship detection processing of the payment user and the associated user based on interaction data of the payment user and the associated user; the relationship detection processing comprising interaction type detection, interaction attenuation calculation and relationship release prediction; a marking processing module configured to perform marking processing on the protocol payment channel and issue a payment channel set carrying the protocol payment channel to the user terminal if the relationship detection fails; and a payment confirmation module configured to perform payment confirmation of the protocol payment channel for the associated user according to a payment processing request of the protocol payment channel sent by the user terminal.

[0009] One or more embodiments of the present specification provide a payment detection processing device, comprising: a processor; and a memory configured to store computer executable instructions which, when executed, cause the processor to: obtain a protocol payment channel opened by an associated user to a payment user; perform interaction type detection based on first interaction data of the payment user and the associated user; if the detection fails, perform interaction attenuation calculation based on second interaction data of the payment user and the associated user; in the case that the obtained interaction attenuation value is less than an attenuation threshold, call a relationship prediction model to perform relationship release prediction of the payment user and the associated user; and if it is determined that the relationship is released based on the prediction result, perform marking processing on the protocol payment channel.

[0010] One or more embodiments of the present specification provide a second payment detection processing device, comprising: a processor; and a memory configured to store computer executable instructions which, when executed, cause the processor to: submit a payment access request to a server based on a payment operation of the payment user; receive and display a payment channel set containing a protocol payment channel issued by the server; the protocol payment channel is marked after performing interaction type detection, interaction attenuation calculation and relationship release prediction on the payment user and an associated user opening the protocol payment channel; and if a payment instruction for the protocol payment channel is detected, send a payment processing request to the server to perform payment confirmation of the protocol payment channel for the associated user.

[0011] One or more embodiments of the present specification provide a third payment detection processing device, comprising: a processor; and a memory configured to store computer executable instructions that, when executed, cause the processor to: obtain a protocol payment channel opened by an associated user to a payment user according to a payment access request submitted by a user terminal of the payment user. Perform relationship detection processing of the payment user and the associated user based on interaction data of the payment user and the associated user. The relationship detection processing includes interaction type detection, interaction decay calculation, and relationship release prediction. If the relationship detection fails, mark the protocol payment channel, and issue a payment channel set carrying the protocol payment channel to the user terminal. According to the payment processing request of the protocol payment channel sent by the user terminal, perform payment confirmation of the protocol payment channel for the associated user.

[0012] One or more embodiments of the present specification provide a storage medium for storing computer executable instructions, which, when executed by a processor, implement the following processes: obtaining a protocol payment channel opened by an associated user to a payment user. Perform interaction type detection based on first interaction data of the payment user and the associated user. If the detection fails, perform interaction decay calculation based on second interaction data of the payment user and the associated user. If the calculated interaction decay value is less than the decay threshold, call a relationship prediction model to perform relationship release prediction of the payment user and the associated user. If it is determined that the relationship is released based on the prediction result, mark the protocol payment channel.

[0013] One or more embodiments of the present specification provide a second storage medium for storing computer executable instructions, which, when executed by a processor, implement the following processes: submitting a payment access request to a server based on a payment operation of the payment user. Receive and display the payment channel set containing the protocol payment channel issued by the server. The protocol payment channel is marked after interaction type detection, interaction decay calculation, and relationship release prediction of the payment user and an associated user who opened the protocol payment channel. If a payment instruction for the protocol payment channel is detected, send a payment processing request to the server to perform payment confirmation of the protocol payment channel for the associated user.

[0014] The one or more embodiments of the specification provide a third storage medium for storing computer executable instructions, which, when executed by a processor, implement the following process: according to a payment access request submitted by a user terminal of a payment user, obtaining a protocol payment channel opened by an associated user to the payment user. Based on the interaction data of the payment user and the associated user, the relationship detection processing of the payment user and the associated user is carried out. The relationship detection processing includes interaction type detection, interaction attenuation calculation and relationship release prediction. If the relationship detection fails, the protocol payment channel is marked, and a payment channel set carrying the protocol payment channel is issued to the user terminal. According to the payment processing request of the protocol payment channel sent by the user terminal, the payment confirmation of the protocol payment channel for the associated user is carried out. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the one or more embodiments of the specification or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the specification, and those skilled in the art can also obtain other drawings according to these drawings without creative labor;

[0016] Figure 1 A payment detection processing method processing flowchart is provided for the one or more embodiments of the specification;

[0017] Figure 2 An interaction type detection processing flowchart is provided for the one or more embodiments of the specification;

[0018] Figure 3 A schematic diagram of the graph structure interaction data is provided for the one or more embodiments of the specification;

[0019] Figure 4 An interaction attenuation calculation processing flowchart is provided for the one or more embodiments of the specification;

[0020] Figure 5 A relationship release prediction processing flowchart is provided for the one or more embodiments of the specification;

[0021] Figure 6 A schematic diagram of the model framework of the relationship prediction model is provided for the one or more embodiments of the specification;

[0022] Figure 7 A schematic diagram of the payment channel list is provided for the one or more embodiments of the specification;

[0023] Figure 8An application payment page provided by one or more embodiments of the present specification;

[0024] Figure 9 An application payment confirmation page provided by one or more embodiments of the present specification;

[0025] Figure 10 A second payment detection processing method processing flowchart provided by one or more embodiments of the present specification;

[0026] Figure 11 A payment detection processing method processing flowchart applied to a close payment scenario provided by one or more embodiments of the present specification;

[0027] Figure 12 A third payment detection processing method processing flowchart provided by one or more embodiments of the present specification;

[0028] Figure 13 A payment detection processing device schematic diagram provided by one or more embodiments of the present specification;

[0029] Figure 14 A second payment detection processing device schematic diagram provided by one or more embodiments of the present specification;

[0030] Figure 15 A third payment detection processing device schematic diagram provided by one or more embodiments of the present specification;

[0031] Figure 16 A payment detection processing device structure schematic diagram provided by one or more embodiments of the present specification;

[0032] Figure 17 A second payment detection processing device structure schematic diagram provided by one or more embodiments of the present specification;

[0033] Figure 18 A third payment detection processing device structure schematic diagram provided by one or more embodiments of the present specification. DETAILED DESCRIPTION

[0034] In order to make the person skilled in the art better understand the technical scheme in one or more embodiments of the present specification, the technical scheme in one or more embodiments of the present specification will be described clearly and completely in conjunction with the drawings in one or more embodiments of the present specification. Obviously, the described embodiments are only part of the embodiments of the present specification, not all. Based on one or more embodiments of the present specification, all other embodiments obtained by those skilled in the art without creative labor should belong to the protection scope of the present document.

[0035] An embodiment of a payment detection processing method provided in the specification is as follows:

[0036] In actual scenarios, users open an agreement payment channel by signing an agreement. The user on the opening side can use the corresponding payment channel of the user on the opening side to make a payment, which greatly facilitates the user's payment. In particular, for users who have no payment ability or have relatively low payment ability, they do not need to perform complex payment recharge, payment channel configuration, and other operations, and can use the agreement payment opened by family members or other users to make a convenient and fast payment. The users who open the agreement payment are often users who have a certain relationship in reality. Through the constraint of the real relationship, the payment security can be guaranteed and payment disputes can be avoided. However, the real relationship between users may change or even break down. In this case, the possibility of payment disputes increases, and the payment security of the agreement payment is also challenged.

[0037] The payment detection processing method provided in the embodiment detects the relationship between the payment user and the associated user before the payment user uses the agreement payment channel to make a payment, based on the associated user opening an agreement payment channel for the payment user. Specifically, first, the interaction type of the payment user and the associated user is detected. If the detection fails, the interaction attenuation of the payment user and the associated user is further calculated. If the calculated interaction attenuation value is less than the attenuation threshold, the relationship between the payment user and the associated user is further predicted. According to the prediction result, it is determined whether the agreement payment channel can be used. If it is determined that the relationship between the payment user and the associated user is removed based on the prediction result, the agreement payment channel is marked for processing. In this way, the relationship between the payment user and the associated user is described in three aspects, which is more in-depth and comprehensive. The relationship between the payment user and the associated user is efficiently and accurately identified. The probability of loss caused by the use of the agreement payment channel when the relationship between the users changes or is removed is reduced, and the security of the payment between the users using the agreement payment channel is improved.

[0038] In step S102, the agreement payment channel opened by the associated user for the payment user is obtained.

[0039] The payment user includes the user on the opening side of the agreement payment channel, which can also be referred to as the target user. The associated user includes the user on the opening side of the agreement payment channel for the payment user. The associated user can be a specific user who has an association relationship with the payment user, such as a relative user who has a relative relationship with the payment user or a social user who has a social relationship with the payment user.

[0040] The protocol payment channel refers to a payment channel opened to the payment user based on the agreement signed by the associated user and the payment user. The protocol payment channel can be a protocol specified payment channel. It should be noted that in the case where the associated user opens the protocol payment channel to the payment user by signing an agreement with the payment user, the protocol payment channel is a payment channel presented to the payment user side. The payment user can use the protocol payment channel for payment processing, but the actual funds or resources generated by using the protocol payment channel are borne by the associated user, that is, the protocol payment channel is bound to the payment account of the associated user.

[0041] In specific implementation, the protocol payment channel opened by the associated user to the payment user can be acquired in the payment processing process of the payment user. For example, in the process of initiating operation by the payment user, the payment channel list of the payment user itself needs to be returned to the payment user. Before returning the payment channel list, the protocol payment channel of the payment user is detected. The protocol payment channel in the payment channel list can be marked according to the detection result. Finally, the payment channel list containing the marked protocol payment channel is returned to the payment user. In addition, the protocol payment channel of the payment user can also be detected at a preset detection period.

[0042] It should be noted that the payment detection process of one protocol payment channel is taken as an example for description in this embodiment. If the payment user has multiple protocol payment channels, the payment detection process can be performed on each protocol payment channel respectively. The payment detection process of the protocol payment channel provided in this embodiment can be referred to, and this embodiment will not be described again.

[0043] In the process of detecting the multiple protocol payment channels of the payment user, the multiple protocol payment channels can be detected in batch processing mode, or each protocol payment channel can be detected in sequence. For example, each protocol payment channel is detected in sequence according to the payment priority of the protocol payment channel, and the detection order is from high to low according to the payment priority. In addition, in order to reduce the detection time of the payment detection process of the multiple protocol payment channels and improve the detection response efficiency of the protocol payment channel, multiple thread processing modes can also be used. Each protocol payment channel is allocated a thread to realize fast detection response of the protocol payment channel.

[0044] In step S104, the interaction type detection is performed based on the first interaction data of the payment user and the associated user.

[0045] In this embodiment, on the basis of the first interaction data of the payment user and the associated user, the interaction type between the payment user and the associated user is detected by detecting the interaction type of the payment user and the associated user. Specifically, it can be detected whether the interaction between the payment user and the associated user is negative interaction or positive interaction, so as to quickly identify the payment user and the associated user with positive interaction. For the payment user and the associated user with negative interaction, in-depth and comprehensive detection can be performed through subsequent interaction attenuation calculation and relationship release prediction.

[0046] Here, the negative interaction refers to an interaction behavior representing that the payment user and the associated user have an abnormal relationship, such as the interaction behavior of deleting a friend by the payment user and the associated user. The positive interaction refers to an interaction behavior representing that the payment user and the associated user have a normal relationship, such as the interaction behavior of adding a friend by the payment user and the associated user.

[0047] In specific implementation, by configuring the interaction type as a positive interaction type or a negative interaction type, the interaction relationship of the payment user and the associated user can be quickly identified, and the rapid response of the interaction type detection is realized. Specifically, in an optional implementation provided by the embodiment, the interaction type detection based on the first interaction data of the payment user and the associated user includes:

[0048] Detecting whether the interaction type carried by the first interaction data is a negative interaction type;

[0049] If the interaction type is a negative interaction type, marking the protocol payment channel;

[0050] If the interaction type is a positive interaction type, detecting whether the generation duration of the first interaction data is greater than a preset time threshold. If yes, it is confirmed that the detection fails. If no, no processing is performed, or the protocol payment channel is marked as a normal payment state.

[0051] In the specific execution process, in order to improve the effectiveness of the interaction type detection, the first interaction data can select one or more latest interaction records in the interaction record of the payment user and the associated user as the first interaction data. In this way, the real-time and effectiveness of the interaction type detection based on the first interaction data are improved on the basis of improving the real-time of the first interaction data.

[0052] For example, after user B and user A sign an intimate payment protocol, user A can use the intimate payment channel to pay based on the intimate payment protocol after the protocol takes effect. The payment fund generated by user A using the intimate payment channel is paid by user B and deducted from the fund account of user B. During the process of user A using the intimate payment channel to pay, the interaction type detection for user A and user B needs to be performed, such as Figure 2The processing procedure of the interaction type detection shown is as follows:

[0053] Step S202, querying the latest interaction message after user A and user B sign the intimate payment agreement;

[0054] Step S204, querying the interaction type of the interaction message;

[0055] Step S206, judging whether the interaction type is a negative interaction type;

[0056] If yes, step S208 is executed;

[0057] If no, steps S210 and S212 are executed;

[0058] Step S208, determining that the intimate payment channel has payment risk;

[0059] In the case that the intimate payment channel has payment risk, user A needs to confirm user B before making payment by using the intimate payment channel;

[0060] Step S210, calculating the interval days between the interaction time of the interaction message and the current time;

[0061] Step S212, judging whether the interval days are less than a specified day threshold;

[0062] If yes, step S214 is executed;

[0063] If no, step S216 is executed;

[0064] Step S214, determining that the intimate payment channel does not have payment risk;

[0065] Step S216, entering the next link of the processing of interaction decay calculation.

[0066] Optionally, the first interaction data includes graph structure interaction data stored in a graph database; wherein, a data node in the graph structure interaction data corresponds to a user identifier of the payment user and a user identifier of the associated user; a data connection in the graph structure interaction data corresponds to a data sequence composed of the user identifier of the payment user, the user identifier of the associated user, an interaction type and / or time information.

[0067] For example, Figure 3 The graph structure interaction data shown is composed of nodes and edges, the node ID is the user ID, and the starting node ID, the target node ID, the interaction type (TYPE) and the timestamp (TIME) uniquely represent an edge, wherein, Figure 3The leftmost edge in the middle can be represented as: 2088xxx03-Add friend-2022 / 09 / 12 10:00:00-2088xxx01.

[0068] In step S106, if the detection fails, interaction attenuation calculation is performed based on second interaction data of the payment user and the associated user.

[0069] In actual scenarios, in the case that the interaction relationship type between the payment user and the associated user is a negative interaction type, that is, after the relationship between the payment user and the associated user is abnormal or broken, the interaction between the payment user and the associated user may become less and less, in other words, the less the interaction between the payment user and the associated user, the more likely the relationship between the payment user and the associated user is abnormal or broken; on the other hand, the interaction frequency of the payment user and the associated user is positively correlated with the activity of the two parties, that is, the interaction frequency attenuation of a high active user after the relationship is abnormal or broken is more obvious than that of a low active user after the relationship is abnormal or broken.

[0070] The embodiment detects the possibility of abnormality or rupture of the relationship between the payment user and the associated user from the attenuation degree of the interaction frequency between the payment user and the associated user, where the interaction attenuation calculation refers to calculating the attenuation degree of the interaction frequency between the payment user and the associated user to depict the possibility of abnormality or rupture of the relationship between the payment user and the associated user.

[0071] In an optional implementation provided by the embodiment, the interaction attenuation calculation based on the second interaction data of the payment user and the associated user includes:

[0072] Based on the total interaction times contained in the second interaction data, the cumulative interaction frequency of the payment user and the associated user is calculated.

[0073] Based on the average activity and interaction interval length of the payment user and the associated user contained in the second interaction data, a time attenuation score is calculated.

[0074] The interaction attenuation value is calculated according to the cumulative interaction frequency and the time attenuation score.

[0075] The second interaction data can be the interaction data of the payment user and the associated user in a specific time interval in the past.

[0076] Further, after the interaction decay calculation based on the second interaction data of the payment user and the associated user, if the interaction decay value obtained by the calculation is greater than or equal to the decay threshold, it indicates that the interaction decay degree between the payment user and the associated user is large. In this case, the protocol payment channel is optionally marked.

[0077] In the above example, the processing process of the interaction decay calculation of user A and user B is as shown in Figure 4

[0078] Step S402, query the total interaction times cnt of user A and user B in the past year;

[0079] Step S404, query the interval days t of the interaction time of the latest interaction message of user A and user B and the current time;

[0080] Step S406, calculate the average value avg of the monthly active days of user A and user B;

[0081] Step S408, calculate the interaction decay score S of user A and user B through the interaction decay scoring formula;

[0082] The interaction decay scoring formula is as follows:

[0083] S = ((2 / (1+e -a*cnt ))-1)*(1-(-f*t / e b*avg*avg-c*avg+d ))

[0084] Wherein, a, b, c, d, f are pre-set constants, cnt is the total interaction times, avg is the average value of the monthly active days of user A and user B, and t is the interval days of the interaction time of the latest interaction message of user A and user B and the current time;

[0085] Step S410, determine whether the interaction decay score S exceeds the preset score threshold;

[0086] If yes, execute step S412;

[0087] If no, execute step S414;

[0088] Step S412, determine that the payment channel of close payment has payment risk;

[0089] In the case that the payment channel of close payment has payment risk, user A needs to be confirmed by user B before payment when using the payment channel of close payment for payment;

[0090] Step S414, enter the next link relationship release prediction processing.

[0091] ​Step S108, in the case that the interaction attenuation value obtained by calculation is less than the attenuation threshold, calling a relationship prediction model to perform relationship removal prediction of the payment user and the associated user.

[0092] In the case that the interaction attenuation value obtained by the above interaction attenuation calculation is less than the attenuation threshold, it indicates that the attenuation degree of interaction between the payment user and the associated user is small, and for this case, the relationship removal prediction of the payment user and the associated user is performed by calling the relationship prediction model, so as to further detect the relationship between the payment user and the associated user. The relationship removal prediction refers to predicting the probability of removal of the user relationship between the payment user and the associated user through detection of the user relationship with uniqueness or exclusivity.

[0093] In the specific detection process, starting from the relationship between the payment user and the associated user, it is detected whether the payment user and the associated user have a user relationship conflicting with the current relationship, so as to detect the abnormality or rupture of the current relationship between the payment user and the associated user; for example, for the user relationship of "lover" and "spouse", since such user relationship has uniqueness or exclusivity, in the case that the current user relationship between the payment user and the associated user is "lover" relationship or "spouse" relationship, if it is detected that the payment user establishes "lover" relationship or "spouse" relationship with other users, it indicates that the "lover" relationship or "spouse" relationship between the payment user and the associated user is likely to have been removed or ruptured, or if it is detected that the associated user establishes "lover" relationship or "spouse" relationship with other users, it indicates that the "lover" relationship or "spouse" relationship between the payment user and the associated user is also likely to have been removed or ruptured.

[0094] In an optional implementation provided by the embodiment, calling the relationship prediction model to perform the relationship removal prediction of the payment user and the associated user includes:

[0095] obtaining first user data of a first associated user set of the payment user and second user data of a second associated user set of the associated user;

[0096] inputting the first user data and the second user data into the relationship prediction model, and predicting a relationship removal value of the payment user and the associated user as the prediction result.

[0097] Further, in the process of predicting the relationship removal value of the payment user and the associated user, a neural network is deployed in the relationship prediction model to improve the efficiency and accuracy of the relationship removal value prediction, and optionally, the relationship prediction model includes a first neural network, a second neural network and a third neural network.

[0098] To further improve the accuracy of the relationship dissolution value prediction, the attention mechanism is introduced for relationship dissolution value calculation. Specifically, in an optional embodiment provided by the present embodiment, the prediction of the relationship dissolution value between the payment user and the associated user comprises:

[0099] The first spliced vector in the vector set of the payment user and the second spliced vector in the vector set of the associated user are respectively input into the corresponding first neural network for vector transformation to obtain a transformed vector;

[0100] The attention weight of the transformed vector is calculated through the second neural network, and the normalization weight is calculated according to the calculated attention weight;

[0101] The weighted vector is calculated according to the normalization weight and the transformed vector, and the weighted vector is input into the third neural network for calculation to obtain the relationship dissolution value.

[0102] Optionally, the attention weight comprises a first attention weight of the payment user and a candidate associated user of the payment user, and a second attention weight of the associated user and a candidate associated user of the associated user. On this basis, in an optional embodiment provided by the present embodiment, the calculation of the normalization weight according to the calculated attention weight comprises: calculating the weight sum of the first attention weight and the second attention weight; calculating the ratio of the first attention weight and the weight sum as the normalization weight.

[0103] On the basis of the above-provided optional embodiment for predicting the relationship dissolution value between the payment user and the associated user, in order to reduce the calculation amount required for vector calculation and improve the calculation efficiency, the present embodiment reduces the calculation amount required for vector calculation by mapping high-dimensional vectors to low-dimensional vectors. Specifically, in an optional embodiment provided by the present embodiment, before the first spliced vector in the vector set of the payment user and the second spliced vector in the vector set of the associated user are respectively input into the corresponding first neural network for vector transformation, the following vector mapping operation is performed:

[0104] The initial feature vector and the initial relationship vector of the payment user are mapped to obtain a first feature mapping vector and a first relationship mapping vector;

[0105] The first feature mapping vector and the first relationship mapping vector are spliced to obtain the first spliced vector;

[0106] And / or,

[0107] vector mapping is performed on the initial feature vector and the initial relationship vector of the associated user to obtain a second feature mapping vector and a second relationship mapping vector;

[0108] vector splicing is performed on the second feature mapping vector and the second relationship mapping vector to obtain the second splicing vector.

[0109] Specifically, in the process of determining the initial feature vector of the payment user and the initial feature vector of the associated user, the feature vectors of the payment user, the associated user, and the candidate associated user of the payment user are fused, and the feature vectors of the payment user, the associated user, and the candidate associated user of the associated user are fused, so as to calculate on the basis of two fusion vectors, convert the conflict calculation of the user relationship into a binary classification problem, and thus realize the detection of the user relationship with uniqueness or exclusivity. In an optional implementation provided by the embodiment, the initial feature vector of the payment user is obtained in the following manner: a first feature vector of the payment user, a second feature vector of the associated user, and a third feature vector of the candidate associated user of the payment user are obtained; the first feature vector, the second feature vector, and the third feature vector are spliced to obtain the initial feature vector of the payment user.

[0110] In an optional implementation provided by the embodiment, the initial relationship vector of the payment user is obtained in the following manner: a first interaction relationship vector between the payment user and the associated user, and a second interaction relationship vector between the payment user and the candidate associated user are obtained; the first interaction relationship vector and the second interaction relationship vector are spliced to obtain the initial relationship vector of the payment user.

[0111] The above provides the determination process of the initial feature vector and the initial relationship vector of the payment user, and the determination process of the initial feature vector and the initial relationship vector of the associated user is similar. Specifically, in an optional implementation provided by the embodiment, the initial feature vector of the associated user is obtained in the following manner: a first feature vector of the payment user, a second feature vector of the associated user, and a fourth feature vector of the candidate associated user of the associated user are obtained; the first feature vector, the second feature vector, and the fourth feature vector are spliced to obtain the initial feature vector of the associated user.

[0112] In an optional implementation of the embodiment, the initial relationship vector of the associated user is obtained in the following manner: a first interaction relationship vector of the payment user and the associated user is obtained, and a third interaction relationship vector of the associated user and a candidate associated user of the associated user is obtained; the first interaction relationship vector and the third interaction relationship vector are spliced to obtain the initial relationship vector of the associated user.

[0113] Optionally, the candidate associated user of the payment user is obtained by screening the interaction user set of the payment user according to the relationship label of the payment user and the associated user; specifically, the interaction user whose relationship label is the same as the relationship label of the payment user and the associated user in the interaction user set of the payment user is screened as the candidate associated user of the payment user.

[0114] Similarly, the candidate associated user of the associated user can also be obtained by screening the interaction user set of the associated user according to the relationship label of the payment user and the associated user; specifically, the interaction user whose relationship label is the same as the relationship label of the payment user and the associated user in the interaction user set of the associated user is screened as the candidate associated user of the associated user.

[0115] The relationship label of the payment user and the associated user can be determined according to the relationship keywords of the two configured by the payment user or the associated user, or the user data of the payment user and the user data of the associated user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the two.

[0116] Based on similar logic, the relationship label of the payment user and the interaction user in the interaction user set of the payment user can be determined according to the relationship keywords of the two configured by the payment user or the interaction user, or the user data of the payment user and the user data of the interaction user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the two; similarly, the relationship label of the associated user and the interaction user in the interaction user set of the associated user can be determined according to the relationship keywords of the two configured by the associated user or the interaction user, or the user data of the associated user and the user data of the interaction user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the two.

[0117] In an optional implementation of the embodiment, if it is determined that the relationship between the payment user and the associated user is not removed based on the prediction result after the relationship prediction model is called to predict the relationship removal between the payment user and the associated user, the payment state of the agreement payment channel is confirmed to be normal.

[0118] In the process of determining whether the relationship between the payment user and the associated user is released based on the prediction result, the relationship release value output by the relationship prediction model can be used to determine whether the relationship release value is greater than the pre-set relationship release threshold. If it is less than or equal to the relationship release threshold, it indicates that the relationship between the payment user and the associated user is abnormal or less likely to be broken, that is, it is determined based on the prediction result that the relationship between the payment user and the associated user is not released; if it is greater than the relationship release threshold, it indicates that the relationship between the payment user and the associated user is abnormal or more likely to be broken, that is, it is determined based on the prediction result that the relationship between the payment user and the associated user is released.

[0119] In the above example, the process of relationship release prediction for user A and user B is as shown in Figure 5

[0120] Step S502, identify whether the close relationship between user A and user B is a marital relationship;

[0121] If yes, execute step S504;

[0122] If no, execute steps S506 to S512;

[0123] Step S504, determine that the close payment channel does not have payment risk;

[0124] Step S506, recall a user set z u that may have a marital relationship with user B;

[0125] Step S508, recall a user set z v that may have a marital relationship with user A;

[0126] Step S510, predict a relationship release prediction score of user A and user B through a relationship prediction model;

[0127] Step S512, determine whether the relationship release prediction score is greater than a relationship release threshold;

[0128] If yes, execute step S514;

[0129] If no, execute step S504;

[0130] Step S514, determine that the close payment channel has payment risk.

[0131] For the relationship pair <u, v> of user B and user A, and the k1th user in the user set z u that may have a marital relationship with user B, ​Extract the feature vectors of the three in the user dimension, such as the user's birth date, asset information, consumption preferences, etc., and splice the feature vectors of the three to obtain In addition, the relationship vector representing the user relationship between user B and user A is extracted, and the relationship vector representing the user relationship between user B and user Splicing the two relationship vectors can obtain

[0132] Similarly, for n users who may have a marital relationship with user B, the feature vector set and the relationship vector set Similarly, for m users who may have a marital relationship with user A, the feature vector set and the relationship vector set

[0133] Map the feature vectors in the two feature vector sets obtained above and the relationship vectors in the two relationship vector sets to low-dimensional vector spaces, respectively, and splice the mapped feature vectors of user B with the relationship vectors to obtain a splice vector set composed of splice vectors and splice the mapped feature vectors of user A with the relationship vectors to obtain a splice vector set composed of splice vectors

[0134] As shown in the model framework of the relationship prediction model Figure 6 , a fully connected neural network is added for each splice vector in the two splice vector sets to perform vector transformation:

[0135] and

[0136] wherein and represent the weight matrix and bias of user B and the i-th user who may have a marital relationship with user B, and finally two sets of transformed vectors are obtained:

[0137]

[0138] Considering that different users who may have a marital relationship with user B have different contributions to the current relationship dissolution prediction, they cannot be treated equally, so an attention mechanism is introduced to learn the attention weight of each user who may have a marital relationship with user B. The attention mechanism is implemented by using a two-layer fully connected neural network:

[0139] and

[0140] wherein, W 1 , W2 , W 3 , W 4 and b 1 , b 2 , b 3 , b 4 are parameters obtained in advance by model training;

[0141] and can be obtained by merging and the attention weight is calculated according to the following formula:

[0142]

[0143] After the attention weight is obtained by calculation, the overall normalization weight can be calculated according to β u-＞v

[0144]

[0145] wherein,

[0146] Finally, the prediction score is calculated by a single-layer fully connected neural network

[0147]

[0148] wherein, W and b represent the weight matrix and the bias, respectively.

[0149] Step S110, if it is determined that the relationship is terminated based on the prediction result, the protocol payment channel is marked for processing.

[0150] In the embodiment, by marking the protocol payment channel for processing, the payment user can perceive that the relationship between the associated user who opens the protocol payment channel and the payment user is abnormal or terminated, thereby improving the payment success rate of the payment user; for the associated user who opens the protocol payment channel for the payment user, the probability of causing capital loss due to the use of the protocol payment channel under the condition that the user relationship changes or is terminated can be reduced, and the safety of the associated user's funds can be ensured.

[0151] In an optional embodiment provided by the embodiment, the protocol payment channel is marked for processing, including:

[0152] marking the protocol payment channel as an abnormal payment state; and / or, based on the abnormal payment state, degrading the payment priority of the protocol payment channel.

[0153] ​In specific implementation, in order to further improve the payment success rate of the payment user, on the basis of reducing the probability of occurrence of capital loss caused by the use of the protocol payment channel in the case of change or termination of the relationship of the user, and ensuring the safety of the funds of the associated user, the payment instruction of the payment user is processed by the payment confirmation mode to the associated user. Specifically, in an optional implementation provided by the embodiment, after the protocol payment channel is marked, the payment confirmation is performed in the following manner:

[0154] According to the payment instruction of the protocol payment channel, a payment confirmation message is sent to the associated user.

[0155] If the confirmation instruction of the payment confirmation message is detected, payment processing is performed based on the protocol payment channel.

[0156] For example, in the payment channel list of user A, the payment priority of the close payment payment channel with user B is adjusted from the first to the last. After adjustment, the payment channel list of user A is as shown in Figure 7 At the same time, the close payment payment channel displays an abnormal prompt "there is a payment risk, and the payment needs to be confirmed", and an access interface "apply for payment confirmation to TA" for applying to user B for confirmation of the current payment is configured. If user A clicks the access interface "apply for payment confirmation to TA", an application payment page as shown in Figure 8 will be generated. After user A clicks the "confirm" button, a confirmation payment prompt message will be sent to user B. After user B triggers the confirmation payment prompt message, a payment confirmation page as shown in Figure 9 is entered.

[0157] It should be noted that the payment detection processing method provided by the present application simultaneously uses the three processing methods of interaction type detection, interaction attenuation calculation and relationship termination prediction for relationship detection processing in the process of relationship detection processing of the payment user and the associated user. In addition, in the specific execution process, any one or any two of the three processing methods of interaction type detection, interaction attenuation calculation and relationship termination prediction can be selected for relationship detection processing according to the specific execution needs. The execution order of the three processing methods of interaction type detection, interaction attenuation calculation and relationship termination prediction in the relationship detection processing process can also be adjusted accordingly, and the three processing methods after the order adjustment are used for relationship detection processing. Similarly, the execution order of any two processing methods in the three processing methods of interaction type detection, interaction attenuation calculation and relationship termination prediction can also be adjusted accordingly in the process of relationship detection processing, and the relationship detection processing is performed by using the two processing methods after the order adjustment.

[0158] The following respectively provide an implementation mode of relationship detection processing by selecting any one of the three processing modes of interaction type detection, interaction attenuation calculation and relationship resolution prediction, an implementation mode of relationship detection processing by selecting any two of the three processing modes of interaction type detection, interaction attenuation calculation and relationship resolution prediction, an implementation mode of relationship detection processing after adjusting the execution order of the three processing modes of interaction type detection, interaction attenuation calculation and relationship resolution prediction, an implementation mode of relationship detection processing after selecting any two of the three processing modes of interaction type detection, interaction attenuation calculation and relationship resolution prediction and adjusting the execution order of the any two, and other implementation modes can be known from the following four implementation mode examples, and the embodiment will not be described here.

[0159] (1) An implementation mode of relationship detection processing by selecting any one of the three processing modes of interaction type detection, interaction attenuation calculation and relationship resolution prediction

[0160] Obtaining a protocol payment channel opened by an associated user to a payment user;

[0161] Performing interaction type detection based on first interaction data of the payment user and the associated user;

[0162] If the detection fails, performing marking processing on the protocol payment channel;

[0163] Alternatively,

[0164] Obtaining a protocol payment channel opened by an associated user to a payment user;

[0165] Performing interaction attenuation calculation based on second interaction data of the payment user and the associated user;

[0166] If the obtained interaction attenuation value is less than the attenuation threshold, performing marking processing on the protocol payment channel.

[0167] Alternatively,

[0168] Obtaining a protocol payment channel opened by an associated user to a payment user;

[0169] Calling a relationship prediction model to perform relationship resolution prediction of the payment user and the associated user;

[0170] If it is determined that the relationship is resolved based on the prediction result, performing marking processing on the protocol payment channel.

[0171] (2) An implementation mode of relationship detection processing by selecting any two of the three processing modes of interaction type detection, interaction attenuation calculation and relationship resolution prediction

[0172] Obtaining a protocol payment channel opened by an associated user to a payment user;

[0173] performing interaction type detection based on first interaction data of the payment user and the associated user;

[0174] If the detection fails, performing interaction attenuation calculation based on second interaction data of the payment user and the associated user;

[0175] In the case where the obtained interaction attenuation value is less than the attenuation threshold, performing marking processing on the agreement payment channel.

[0176] Alternatively,

[0177] Obtaining an agreement payment channel opened by an associated user to a payment user;

[0178] performing interaction type detection based on first interaction data of the payment user and the associated user;

[0179] If the detection fails, calling a relationship prediction model to perform relationship release prediction of the payment user and the associated user;

[0180] If it is determined that the relationship is released based on the prediction result, performing marking processing on the agreement payment channel.

[0181] Alternatively,

[0182] Obtaining an agreement payment channel opened by an associated user to a payment user;

[0183] performing interaction attenuation calculation based on second interaction data of the payment user and the associated user;

[0184] In the case where the obtained interaction attenuation value is less than the attenuation threshold, calling a relationship prediction model to perform relationship release prediction of the payment user and the associated user;

[0185] If it is determined that the relationship is released based on the prediction result, performing marking processing on the agreement payment channel.

[0186] (3) The implementation mode of the relationship detection processing after the three processing modes of interaction type detection, interaction attenuation calculation and relationship release prediction are adjusted in execution order

[0187] Obtaining an agreement payment channel opened by an associated user to a payment user;

[0188] performing interaction attenuation calculation based on second interaction data of the payment user and the associated user;

[0189] In the case where the obtained interaction attenuation value is less than the attenuation threshold, performing interaction type detection based on first interaction data of the payment user and the associated user;

[0190] If the detection fails, a relationship prediction model is called to predict the relationship between the payment user and the associated user.

[0191] If it is determined that the relationship is dissolved based on the prediction result, the protocol payment channel is marked.

[0192] (4) An implementation manner in which any two of the three processing manners of interaction type detection, interaction decay calculation, and relationship dissolution prediction are selected and the execution order of the any two is adjusted, and then the relationship detection processing is performed

[0193] An agreement payment channel opened by the associated user to the payment user is obtained.

[0194] Interaction decay calculation is performed based on second interaction data of the payment user and the associated user.

[0195] If the obtained interaction decay value is less than the decay threshold, interaction type detection is performed based on first interaction data of the payment user and the associated user.

[0196] If the detection fails, the protocol payment channel is marked.

[0197] It should be noted that the specific implementation details of the various implementation manners provided herein can be found in the corresponding content in steps S102 to S110 described above, and thus will not be described again in this embodiment.

[0198] A second payment detection processing method embodiment provided in the specification:

[0199] In step S1002, a payment access request is submitted to a server based on a payment operation of the payment user.

[0200] The payment user in this embodiment includes a user who initiates payment and is opened to a protocol payment channel. The other party user opposite to the payment user is an associated user, and the associated user includes a user who opens a protocol payment channel to the payment user. The associated user can be a specific user who has an association relationship with the payment user, such as a relative user who has a relative relationship with the payment user, or a social user who has a social relationship with the payment user. The payment operation can be a payment operation for an online order to be paid, and can also be a payment operation for an offline order.

[0201] In step S1004, a payment channel set containing a protocol payment channel issued by the server is received and displayed.

[0202] The protocol payment channel refers to a payment channel opened for a payment user based on a protocol signed by an associated user and the payment user. The protocol payment channel can be a payment channel specified by the protocol. It should be noted that in the case where the associated user opens the protocol payment channel for the payment user by signing the protocol with the payment user, the protocol payment channel is a payment channel presented to the payment user side. The payment user can use the protocol payment channel for payment processing, but the actual fund or resource outflow generated by using the protocol payment channel is borne by the associated user, that is, the protocol payment channel is bound to the payment account of the associated user.

[0203] The above-mentioned payment access request is submitted to the server, and after the server receives the payment access request, the payment channel set composed of at least one payment channel of the payment user is first obtained. In the case where the payment channel set contains a protocol payment channel, relationship detection processing is performed on the payment user and the associated user who opens the protocol payment channel, and the protocol payment channel in the payment channel list can be marked according to the detection result. Finally, the payment channel list containing the marked protocol payment channel is returned to the payment user. Optionally, the protocol payment channel is marked after the interaction type detection, interaction attenuation calculation and relationship removal prediction of the payment user and the associated user who opens the protocol payment channel.

[0204] In specific implementation, in the process of relationship detection processing of the payment user and the associated user who opens the protocol payment channel by the server, interaction type detection, interaction attenuation calculation and relationship removal prediction can be performed on the payment user and the associated user. The interaction type detection can be based on the first interaction data of the payment user and the associated user, the interaction attenuation calculation can be based on the second interaction data, and the relationship removal prediction can be performed by calling a relationship prediction model. Further, the interaction attenuation calculation can be performed when the detection of the interaction type detection fails, and the relationship removal prediction can be performed when the interaction attenuation value obtained by the interaction attenuation calculation is less than an attenuation threshold.

[0205] In an optional implementation provided by the embodiment, the interaction type detection is implemented in the following manner:

[0206] Detect whether the interaction type carried by the first interaction data of the payment user and the associated user is a negative interaction type;

[0207] If it is a negative interaction type, mark the protocol payment channel;

[0208] If it is a positive interaction type, detect whether the generation time length of the first interaction data is greater than a preset time threshold. If yes, perform the interaction attenuation calculation.

[0209] Here, on the basis of the first interaction data of the payment user and the associated user, the interaction type between the payment user and the associated user is detected by detecting the interaction type of the payment user and the associated user, and specifically, whether the interaction between the payment user and the associated user is negative interaction or positive interaction is detected, so as to quickly identify the payment user and the associated user with positive interaction. For the payment user and the associated user with negative interaction, in-depth and comprehensive detection can be performed through subsequent interaction decay calculation and relationship release prediction.

[0210] Among them, the negative interaction refers to the interaction behavior representing that the relationship between the payment user and the associated user is abnormal, such as the interaction behavior of deleting friends by the payment user and the associated user; the positive interaction refers to the interaction behavior representing that the relationship between the payment user and the associated user is normal, such as the interaction behavior of adding friends by the payment user and the associated user.

[0211] Optionally, the first interaction data includes graph structure interaction data stored in a graph database; wherein the data node in the graph structure interaction data corresponds to the user identifier of the payment user and the user identifier of the associated user; the data connection in the graph structure interaction data corresponds to a data sequence composed of the user identifier of the payment user, the user identifier of the associated user, the interaction type and / or the time information.

[0212] In order to improve the effectiveness of the interaction type detection, the first interaction data can select one or more latest interaction records in the interaction record of the payment user and the associated user as the first interaction data, so as to improve the real-time performance of the first interaction data and the real-time performance and effectiveness of the interaction type detection based on the first interaction data.

[0213] In actual scenarios, in the case that the interaction relationship type between the payment user and the associated user is negative interaction type, that is, after the relationship between the payment user and the associated user is abnormal or broken, the interaction between the payment user and the associated user may become less and less, in other words, the less the interaction between the payment user and the associated user, the greater the possibility that the relationship between the payment user and the associated user is abnormal or broken; on the other hand, the interaction frequency of the payment user and the associated user is positively correlated with the activity of the two parties, that is, the decay of the interaction frequency of the high active user after the relationship is abnormal or broken is more obvious than that of the low active user. This embodiment detects the possibility that the relationship between the payment user and the associated user is abnormal or broken from the decay degree of the interaction frequency of the payment user and the associated user, and the interaction decay calculation refers to calculating the decay degree of the interaction frequency of the payment user and the associated user, so as to depict the possibility that the relationship between the payment user and the associated user is abnormal or broken.

[0214] In an optional implementation of the embodiment, the interaction attenuation calculation is implemented in the following manner:

[0215] Based on the total number of interactions contained in the second interaction data of the payment user and the associated user, the cumulative interaction frequency of the payment user and the associated user is calculated.

[0216] Based on the average activity and interaction interval length of both the payment user and the associated user contained in the second interaction data, a time attenuation score is calculated.

[0217] According to the cumulative interaction frequency and the time attenuation score, an interaction attenuation value is calculated, and the relationship resolution prediction is performed if the interaction attenuation value is less than an attenuation threshold.

[0218] The second interaction data can be the interaction data of the payment user and the associated user within a specific time interval in the past.

[0219] In an optional implementation of the embodiment, the relationship resolution prediction is implemented in the following manner:

[0220] The first splicing vector in the vector set of the payment user and the second splicing vector in the vector set of the associated user are respectively input into corresponding first neural networks for vector transformation to obtain transformed vectors.

[0221] Attention weight calculation of the transformed vectors is performed through a second neural network, and a normalization weight is calculated according to the calculated attention weight.

[0222] A weighted vector is calculated according to the normalization weight and the feature vector, the weighted vector is input into a third neural network for calculation, and the marking process is performed if the obtained relationship resolution value is greater than a preset threshold.

[0223] Optionally, the attention weight includes a first attention weight of the payment user and a candidate associated user of the payment user, and a second attention weight of the associated user and a candidate associated user of the associated user. On this basis, in an optional implementation of the embodiment, the calculation of the normalization weight according to the calculated attention weight includes: calculating the weight sum of the first attention weight and the second attention weight; calculating the ratio of the first attention weight and the weight sum as the normalization weight.

[0224] The above is based on the relationship between the payment user and the associated user, and detects whether the payment user and the associated user exist a user relationship conflicting with the current relationship, so as to detect the abnormality or rupture of the current relationship between the payment user and the associated user; for example, for the user relationship of "lover" and "spouse", since such user relationship exists uniqueness or exclusivity, therefore, in the case that the current user relationship between the payment user and the associated user is "lover" relationship or "spouse" relationship, if it is detected that the payment user establishes "lover" relationship or "spouse" relationship with other user, it indicates that the "lover" relationship or "spouse" relationship between the payment user and the associated user is likely to have been dissolved or ruptured, or, if it is detected that the associated user establishes "lover" relationship or "spouse" relationship with other user, it indicates that the "lover" relationship or "spouse" relationship between the payment user and the associated user is also likely to have been dissolved or ruptured.

[0225] On the basis of the above provided optional implementation of relationship dissolution prediction, in order to reduce the amount of calculation required for vector calculation and improve the calculation efficiency, the embodiment reduces the amount of calculation required for vector calculation by mapping high-dimensional vectors to low-dimensional vectors. Specifically, in an optional implementation provided by the embodiment, before the operation of inputting the first splicing vector in the vector set of the payment user and the second splicing vector in the vector set of the associated user into the corresponding first neural network for vector transformation to obtain a transformed vector, the following vector mapping operation is performed:

[0226] Vector mapping is performed on the initial feature vector and the initial relationship vector of the payment user to obtain a first feature mapping vector and a first relationship mapping vector;

[0227] The first feature mapping vector and the first relationship mapping vector are spliced to obtain the first splicing vector;

[0228] And / or,

[0229] Vector mapping is performed on the initial feature vector and the initial relationship vector of the associated user to obtain a second feature mapping vector and a second relationship mapping vector;

[0230] The second feature mapping vector and the second relationship mapping vector are spliced to obtain the second splicing vector.

[0231] Specifically, in the process of determining the initial feature vector of the payment user and the initial feature vector of the associated user, the feature vectors of the payment user, the associated user and the candidate associated user of the payment user are fused, and the feature vectors of the payment user, the associated user and the candidate associated user of the associated user are fused, so as to calculate on the basis of the two fused vectors, convert the conflict calculation of the user relationship into a binary classification problem, and thus realize the detection of the user relationship with uniqueness or exclusivity. In an optional implementation provided by the embodiment, the initial feature vector of the payment user is obtained in the following manner: obtaining a first feature vector of the payment user, a second feature vector of the associated user and a third feature vector of the candidate associated user of the payment user; performing vector splicing on the first feature vector, the second feature vector and the third feature vector to obtain the initial feature vector of the payment user.

[0232] In an optional implementation provided by the embodiment, the initial relationship vector of the payment user is obtained in the following manner: obtaining a first interaction relationship vector between the payment user and the associated user, and a second interaction relationship vector between the payment user and the candidate associated user; and performing splicing on the first interaction relationship vector and the second interaction relationship vector to obtain the initial relationship vector of the payment user.

[0233] The above provides the determination process of the initial feature vector and the initial relationship vector of the payment user, and the determination process of the initial feature vector and the initial relationship vector of the associated user is similar thereto. Specifically, in an optional implementation provided by the embodiment, the initial feature vector of the associated user is obtained in the following manner: obtaining a first feature vector of the payment user, a second feature vector of the associated user and a fourth feature vector of the candidate associated user of the associated user; and performing vector splicing on the first feature vector, the second feature vector and the fourth feature vector to obtain the initial feature vector of the associated user.

[0234] In addition, in an optional implementation provided by the embodiment, the initial relationship vector of the associated user is obtained in the following manner: obtaining a first interaction relationship vector between the payment user and the associated user, and a third interaction relationship vector between the associated user and the candidate associated user of the associated user; and performing splicing on the first interaction relationship vector and the third interaction relationship vector to obtain the initial relationship vector of the associated user.

[0235] Optionally, the candidate associated user of the payment user is obtained by screening the payment user's interactive user set according to the relationship label of the payment user and the associated user; specifically, the interactive user whose relationship label is the same as the relationship label of the payment user and the associated user in the payment user's interactive user set is screened as the candidate associated user of the payment user.

[0236] Similarly, the candidate associated user of the associated user can also be obtained by screening the associated user's interactive user set according to the relationship label of the payment user and the associated user; specifically, the interactive user whose relationship label is the same as the relationship label of the payment user and the associated user in the associated user's interactive user set is screened as the candidate associated user of the associated user.

[0237] Among them, the relationship label of the payment user and the associated user can be determined according to the relationship keywords configured by the payment user or the associated user, or the user data of the payment user and the user data of the associated user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the payment user and the associated user.

[0238] Based on similar logic, the relationship label of the payment user and the interactive user in the payment user's interactive user set can be determined according to the relationship keywords configured by the payment user or the interactive user, or the user data of the payment user and the user data of the interactive user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the payment user and the interactive user; similarly, the relationship label of the associated user and the interactive user in the associated user's interactive user set can be determined according to the relationship keywords configured by the associated user or the interactive user, or the user data of the associated user and the user data of the interactive user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the associated user and the interactive user.

[0239] It should be noted that in the process of relationship detection and processing of the payment user and the associated user, the three processing methods of interaction type detection, interaction decay calculation and relationship removal prediction can be used simultaneously and sequentially for relationship detection and processing; in addition, in the specific execution process, any one or any two of the three processing methods of interaction type detection, interaction decay calculation and relationship removal prediction can be selected for relationship detection and processing according to the specific execution needs; the execution order of the three processing methods of interaction type detection, interaction decay calculation and relationship removal prediction in the relationship detection and processing process can also be adjusted accordingly, and the three processing methods after the order adjustment are used for relationship detection and processing; similarly, the execution order of any two processing methods in the three processing methods of interaction type detection, interaction decay calculation and relationship removal prediction in the relationship detection and processing process can also be adjusted accordingly, and the two processing methods after the order adjustment are used for relationship detection and processing.

[0240] It should be noted that the processing procedures of the interaction type detection, interaction attenuation calculation and relationship release prediction provided in the embodiment are only illustrative, and the specific processing procedures and execution examples of the interaction type detection, interaction attenuation calculation and relationship release prediction can be provided with reference to the corresponding content provided in the method embodiment steps S104 to S108, and the embodiment will not be described here.

[0241] In the embodiment, by marking the protocol payment channel, the payment user can perceive that the relationship between the associated user who opens the protocol payment channel and the payment user is abnormal or released, thereby improving the payment success rate of the payment user; for the associated user who opens the protocol payment channel to the payment user, the probability of causing capital loss due to the use of the protocol payment channel when the user relationship changes or is released is reduced, and the safety of the associated user's funds is ensured. In an optional implementation provided by the embodiment, the marking processing of the protocol payment channel includes: marking the protocol payment channel as an abnormal payment state; and / or, performing degradation processing on the payment priority of the protocol payment channel based on the abnormal payment state.

[0242] Step S1006, if the payment instruction for the protocol payment channel is detected, a payment processing request is sent to the server to perform payment confirmation of the protocol payment channel for the associated user.

[0243] In specific implementation, on the basis of reducing the probability of causing capital loss due to the use of the protocol payment channel when the user relationship changes or is released, and ensuring the safety of the associated user's funds, in order to further improve the payment success rate of the payment user, the embodiment processes the payment instruction of the payment user by performing payment confirmation to the associated user, specifically, in an optional implementation provided by the embodiment, the payment confirmation of the protocol payment channel for the associated user includes: sending a payment confirmation message to the associated user and receiving a confirmation instruction of the payment confirmation message; based on the confirmation instruction, performing payment processing through the protocol payment channel.

[0244] The implementation process of the second payment detection processing method provided above can be executed by the user terminal, and the implementation process of the third payment detection processing method provided in the following method embodiment can be executed by the server, and the two cooperate with each other in the execution process, therefore, the corresponding content of the third method embodiment can be referred to by reading the above implementation process, and correspondingly, the implementation process of the third method embodiment can also be referred to by reading the corresponding content of the above method embodiment.

[0245] The following takes the application of a payment detection processing method provided in the embodiment in an intimate payment scenario as an example, combined with Figure 11Further description is made to the payment detection processing method provided in the embodiment with reference to Figure 11 The payment detection processing method applied to the intimate payment scenario specifically includes the following steps.

[0246] In step S1102, an order payment operation initiated by user A for order payment is acquired.

[0247] In step S1104, a payment access request of the order payment operation is submitted to a server.

[0248] In step S1118, a payment channel set containing the intimate payment channel after the marking processing is received and displayed.

[0249] In step S1120, a payment instruction for the intimate payment channel is detected.

[0250] In step S1122, a payment processing request of the intimate payment channel is sent to the server.

[0251] The steps S1102 to S1104 and the steps S1118 to S1122 provided in the embodiment are executed by the client. It should be noted that the processing procedures of the steps S1102 to S1104 and the steps S1118 to S1122 executed by the client are matched with the processing procedures of the steps S1106 to S1116 and the steps S1124 to S1126 executed by the server. Therefore, when reading the embodiment, the corresponding content of the steps S1106 to S1116 and the steps S1124 to S1126 is referred to, and vice versa, when reading the steps S1106 to S1116 and the steps S1124 to S1126, the corresponding content of the steps S1102 to S1104 and the steps S1118 to S1122 provided in the embodiment is referred to.

[0252] The third payment detection processing method provided in the specification includes the following steps:

[0253] In step S1202, an agreement payment channel opened by an associated user to a payment user is acquired according to a payment access request submitted by a user terminal of the payment user.

[0254] The payment user in the embodiment includes a user of a party to which the agreement payment channel is opened, which can also be referred to as a target user. The associated user includes a user of a party to which the agreement payment channel is opened to the payment user. The associated user can be a specific user having an association relationship with the payment user, such as a relative user having a relative relationship with the payment user or a social user having a social relationship with the payment user.

[0255] The protocol payment channel refers to a payment channel opened for a payment user based on a protocol signed by the associated user and the payment user. The protocol payment channel can be a protocol-specified payment channel. It should be noted that in the case where the associated user opens the protocol payment channel for the payment user by signing the protocol with the payment user, the protocol payment channel is a payment channel presented to the payment user side. The payment user can use the protocol payment channel for payment processing, but the actual funds or resources generated by using the protocol payment channel are borne by the associated user, that is, the protocol payment channel is bound to the payment account of the associated user.

[0256] In specific implementation, the protocol payment channel opened by the associated user for the payment user can be acquired during the payment processing of the payment user. For example, during the process of initiating an operation by the payment user, the payment user's own payment channel list needs to be returned to the payment user. Before returning the payment channel list, the protocol payment channel of the payment user is detected. The protocol payment channel in the payment channel list can be marked according to the detection result. Finally, the payment channel list containing the marked protocol payment channel is returned to the payment user. In addition, the protocol payment channel of the payment user can also be detected at a preset detection period.

[0257] It should be noted that the payment detection process of one protocol payment channel is taken as an example for illustration in this embodiment. If the payment user has multiple protocol payment channels, the payment detection process can be performed on each protocol payment channel respectively. The payment detection process of the protocol payment channel provided in this embodiment can be referred to, and this embodiment will not be described again.

[0258] During the payment detection process of multiple protocol payment channels of the payment user, the payment detection process of the multiple protocol payment channels can be performed in batch processing mode, or the payment detection process of each protocol payment channel can be performed in sequence. For example, the payment detection process of each protocol payment channel is performed in sequence according to the payment priority of the protocol payment channel and the detection order from high to low according to the payment priority. In addition, in order to reduce the detection time of the payment detection process of the multiple protocol payment channels and improve the detection response efficiency of the protocol payment channel, multiple thread processing modes can also be used. Each protocol payment channel is assigned a thread to achieve fast detection response of the protocol payment channel.

[0259] In step S1204, relationship detection processing of the payment user and the associated user is performed based on interaction data of the payment user and the associated user.

[0260] Optionally, the relationship detection process comprises interaction type detection, interaction attenuation calculation and relationship elimination prediction. In particular, the interaction type detection can be based on first interaction data of the payment user and the associated user, the interaction attenuation calculation can be based on second interaction data, and the relationship elimination prediction can be performed by invoking a relationship prediction model. Further, the interaction attenuation calculation can be performed when the interaction type detection fails, and the relationship elimination prediction can be performed when the interaction attenuation value obtained by the interaction attenuation calculation is less than an attenuation threshold.

[0261] In an optional embodiment provided by the present embodiment, the interaction type detection is implemented in the following manner:

[0262] detecting whether the interaction type carried by the first interaction data of the payment user and the associated user is a negative interaction type;

[0263] if the interaction type is a negative interaction type, marking the protocol payment channel;

[0264] if the interaction type is a positive interaction type, detecting whether the generation duration of the first interaction data is greater than a preset time threshold, and if yes, performing the interaction attenuation calculation.

[0265] Here, based on the first interaction data of the payment user and the associated user, the interaction type between the payment user and the associated user is detected by performing interaction type detection on the payment user and the associated user. Specifically, it can be detected whether the interaction between the payment user and the associated user is a negative interaction or a positive interaction, so as to quickly identify the payment user and the associated user with positive interaction. For the payment user and the associated user with negative interaction, further comprehensive detection can be performed through subsequent interaction attenuation calculation and relationship elimination prediction.

[0266] Here, negative interaction refers to an interaction behavior representing that the payment user and the associated user have an abnormal relationship, such as an interaction behavior of deleting a friend by the payment user and the associated user. Positive interaction refers to an interaction behavior representing that the payment user and the associated user have a normal relationship, such as an interaction behavior of adding a friend by the payment user and the associated user.

[0267] In order to improve the effectiveness of the interaction type detection, the first interaction data can be selected from one or more latest interaction records of the payment user and the associated user as the first interaction data. In this way, the real-time performance and effectiveness of the interaction type detection based on the first interaction data are improved on the basis of improving the real-time performance of the first interaction data.

[0268] For example, after user B signs the close payment agreement with user A, user A can use the close payment channel to make payment based on the close payment agreement, and the payment fund generated by user A using the close payment channel is paid by user B and deducted from the fund account of user B. During the process of user A making payment using the close payment channel, the interaction type detection needs to be performed on user A and user B, such as Figure 2 The processing procedure of the interaction type detection is shown in the following:

[0269] Step S202, querying the latest interaction message after user A and user B sign the close payment agreement;

[0270] Step S204, querying the interaction type of the interaction message;

[0271] Step S206, judging whether the interaction type is a negative interaction type;

[0272] If yes, step S208 is performed;

[0273] If no, steps S210 and S212 are performed;

[0274] Step S208, determining that the close payment channel has payment risk;

[0275] In the case that the close payment channel has payment risk, user A needs to confirm with user B before making payment using the close payment channel;

[0276] Step S210, calculating the interval days between the interaction time of the interaction message and the current time;

[0277] Step S212, judging whether the interval days are less than a specified day threshold;

[0278] If yes, step S214 is performed;

[0279] If no, step S216 is performed;

[0280] Step S214, determining that the close payment channel does not have payment risk;

[0281] Step S216, entering the next link of the processing of interaction decay calculation.

[0282] Optionally, the first interaction data includes graph structure interaction data stored in a graph database; wherein, a data node in the graph structure interaction data corresponds to a user identifier of the payment user and a user identifier of the associated user; a data connection in the graph structure interaction data corresponds to a data sequence composed of the user identifier of the payment user, the user identifier of the associated user, the interaction type and / or the time information.

[0283] For example, Figure 3 The graph structure interaction data shown is composed of nodes and edges, and the node ID is the user ID. The starting node ID, target node ID, interaction type (TYPE), and timestamp (TIME) uniquely represent an edge, where, Figure 3 The leftmost edge in the middle can be represented as: 2088xxx03-Add friend-2022 / 09 / 12 10:00:00-2088xxx01.

[0284] In actual scenarios, when the interaction relationship type between the payment user and the associated user is a negative interaction type, that is, after the relationship between the payment user and the associated user is abnormal or broken, the interaction between the payment user and the associated user may become less and less, in other words, the less the interaction between the payment user and the associated user, the more likely the relationship between the payment user and the associated user is abnormal or broken. On the other hand, the interaction frequency between the payment user and the associated user is positively correlated with the activity of the two parties, that is, the decay of the interaction frequency after the relationship between the high-activity user and the associated user is abnormal or broken is more obvious than that of the low-activity user. This embodiment detects the possibility of abnormal or broken relationship between the payment user and the associated user from the decay degree of the interaction frequency between the payment user and the associated user. Here, the interaction decay calculation refers to calculating the decay degree of the interaction frequency between the payment user and the associated user to depict the possibility of abnormal or broken relationship between the payment user and the associated user.

[0285] In an optional implementation provided by this embodiment, the interaction decay calculation is implemented in the following manner:

[0286] Based on the total number of interactions contained in the second interaction data of the payment user and the associated user, calculate the cumulative interaction frequency of the payment user and the associated user;

[0287] Based on the average activity and interaction interval length of the payment user and the associated user contained in the second interaction data, calculate a time decay score;

[0288] According to the cumulative interaction frequency and the time decay score, calculate an interaction decay value, and perform the relationship resolution prediction when the interaction decay value is less than a decay threshold.

[0289] The second interaction data can be the interaction data of the payment user and the associated user in a specific time interval in the past.

[0290] Following the above example, the processing process of the interaction decay calculation of user A and user B is as followsFigure 4 As shown:

[0291] Step S402, query the total interaction times cnt of user A and user B in the past year;

[0292] Step S404, query the interval days t of the interaction time of the latest interaction message of user A and user B and the current time;

[0293] Step S406, calculate the average value avg of the monthly active days of user A and user B;

[0294] Step S408, calculate the interaction decay score S of user A and user B through the interaction decay scoring formula;

[0295] The interaction decay scoring formula is as follows:

[0296] S = ((2 / (1+e -a*cnt ))-1)*(1-(-f*t / e b*avg*avg-c*avg+d ))

[0297] Wherein, a, b, c, d, f are pre-set constants, cnt is the total interaction times, avg is the average value of the monthly active days of user A and user B, and t is the interval days of the interaction time of the latest interaction message of user A and user B and the current time;

[0298] Step S410, determine whether the interaction decay score S exceeds the preset score threshold;

[0299] If yes, execute step S412;

[0300] If no, execute step S414;

[0301] Step S412, determine that the payment channel of close payment has payment risk;

[0302] In the case that the payment channel of close payment has payment risk, user A needs to be confirmed by user B before payment when using the payment channel of close payment for payment;

[0303] Step S414, enter the next link relationship release prediction processing.

[0304] In an optional embodiment provided by the embodiment, the relationship release prediction is realized in the following way:

[0305] The first splicing vector in the vector set of the payment user and the second splicing vector in the vector set of the associated user are respectively input into the corresponding first neural network for vector transformation to obtain a transformed vector;

[0306] The attention weight calculation of the transformation vector is performed through the second neural network, and a normalization weight is calculated according to the calculated attention weight;

[0307] A weighted vector is calculated according to the normalization weight and the feature vector, the weighted vector is input into the third neural network for calculation, and the marking processing is performed when the obtained relationship resolution value is greater than a preset threshold.

[0308] Optionally, the attention weight includes a first attention weight of the payment user and a candidate associated user of the payment user, and a second attention weight of the associated user and a candidate associated user of the associated user. On this basis, in an optional embodiment provided by the embodiment, the normalization weight is calculated according to the calculated attention weight, including: calculating a weight sum of the first attention weight and the second attention weight; and calculating a ratio of the first attention weight to the weight sum as the normalization weight.

[0309] The above is based on the relationship between the payment user and the associated user, and detects whether the payment user and the associated user exist a user relationship conflicting with the current relationship, so as to detect the abnormality or rupture of the current relationship between the payment user and the associated user; for example, for the user relationship of "lover" and "spouse", since such user relationship exists uniqueness or exclusivity, therefore, in the case that the current user relationship between the payment user and the associated user is "lover" relationship or "spouse" relationship, if it is detected that the payment user establishes "lover" relationship or "spouse" relationship with other user, it indicates that the "lover" relationship or "spouse" relationship between the payment user and the associated user is likely to have been dissolved or ruptured, or if it is detected that the associated user establishes "lover" relationship or "spouse" relationship with other user, it indicates that the "lover" relationship or "spouse" relationship between the payment user and the associated user is also likely to have been dissolved or ruptured.

[0310] On the basis of the optional embodiment of the relationship resolution prediction provided above, in order to reduce the calculation amount required for vector calculation and improve the calculation efficiency, the embodiment reduces the calculation amount required for vector calculation by mapping high-dimensional vectors to low-dimensional vectors. Specifically, in an optional embodiment provided by the embodiment, before the operation of inputting the first splicing vector in the vector set of the payment user and the second splicing vector in the vector set of the associated user into the corresponding first neural network for vector transformation to obtain a transformation vector, the following vector mapping operation is performed:

[0311] The initial feature vector and the initial relationship vector of the payment user are mapped to obtain a first feature mapping vector and a first relationship mapping vector;

[0312] vector splicing of the first feature mapping vector and the first relationship mapping vector to obtain the first spliced vector;

[0313] and / or,

[0314] vector mapping of the initial feature vector and the initial relationship vector of the associated user to obtain a second feature mapping vector and a second relationship mapping vector;

[0315] vector splicing of the second feature mapping vector and the second relationship mapping vector to obtain the second spliced vector.

[0316] Specifically, in the process of determining the initial feature vector of the payment user and the initial feature vector of the associated user, the feature vectors of the payment user, the associated user, and the candidate associated user of the payment user are fused, and the feature vectors of the payment user, the associated user, and the candidate associated user of the associated user are fused, so as to calculate on the basis of the two fused vectors, convert the conflict calculation of the user relationship into a binary classification problem, and thus realize the detection of the user relationship with uniqueness or exclusivity. In an optional implementation provided by the embodiment, the initial feature vector of the payment user is obtained by: obtaining a first feature vector of the payment user, a second feature vector of the associated user, and a third feature vector of the candidate associated user of the payment user; and performing vector splicing on the first feature vector, the second feature vector, and the third feature vector to obtain the initial feature vector of the payment user.

[0317] In an optional implementation provided by the embodiment, the initial relationship vector of the payment user is obtained by: obtaining a first interaction relationship vector between the payment user and the associated user, and a second interaction relationship vector between the payment user and the candidate associated user; and performing splicing on the first interaction relationship vector and the second interaction relationship vector to obtain the initial relationship vector of the payment user.

[0318] The above provides the determination process of the initial feature vector and the initial relationship vector of the payment user, and the determination process of the initial feature vector and the initial relationship vector of the associated user is similar. Specifically, in an optional implementation provided by the embodiment, the initial feature vector of the associated user is obtained by: obtaining a first feature vector of the payment user, a second feature vector of the associated user, and a fourth feature vector of the candidate associated user of the associated user; and performing vector splicing on the first feature vector, the second feature vector, and the fourth feature vector to obtain the initial feature vector of the associated user.

[0319] In an optional implementation of the embodiment, the initial relationship vector of the associated user is obtained by the following method: obtaining a first interaction relationship vector of the payment user and the associated user, and a third interaction relationship vector of the associated user and a candidate associated user of the associated user; and splicing the first interaction relationship vector and the third interaction relationship vector to obtain the initial relationship vector of the associated user.

[0320] Optionally, the candidate associated user of the payment user is obtained by screening the interaction user set of the payment user according to the relationship label of the payment user and the associated user; specifically, the interaction user with the same relationship label as the relationship label of the payment user and the associated user is screened from the interaction user set of the payment user as the candidate associated user of the payment user.

[0321] Similarly, the candidate associated user of the associated user can also be obtained by screening the interaction user set of the associated user according to the relationship label of the payment user and the associated user; specifically, the interaction user with the same relationship label as the relationship label of the payment user and the associated user is screened from the interaction user set of the associated user as the candidate associated user of the associated user.

[0322] The relationship label of the payment user and the associated user can be determined according to the relationship keywords of the two configured by the payment user or the associated user, or the user data of the payment user and the user data of the associated user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the two.

[0323] Based on similar logic, the relationship label of the payment user and the interaction user in the interaction user set of the payment user can be determined according to the relationship keywords of the two configured by the payment user or the interaction user, or the user data of the payment user and the user data of the interaction user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the two; similarly, the relationship label of the associated user and the interaction user in the interaction user set of the associated user can be determined according to the relationship keywords of the two configured by the associated user or the interaction user, or the user data of the associated user and the user data of the interaction user can be input into a relationship detection model for user relationship detection to obtain the relationship label of the two.

[0324] Following the above example, the processing process for relationship resolution prediction of user A and user B is as shown in Figure 5

[0325] Step S502, identifying whether the close relationship between user A and user B is a husband and wife relationship;

[0326] If so, step S504 is performed;

[0327] ​If not, proceed to steps S506 to S512;

[0328] Step S504: Determine that there is no payment risk in the Intimate Payment payment channel;

[0329] Step S506: Recall the set of users z who may have a marital relationship with user B. u ;

[0330] Step S508: Recall the set of users z who may have a marital relationship with user A. v ;

[0331] Step S510: Deprecate the predicted score by predicting the relationship between user A and user B using the relationship prediction model;

[0332] Step S512: Determine whether the relationship termination prediction score is greater than the relationship termination threshold;

[0333] If so, proceed to step S514;

[0334] If not, proceed to step S504;

[0335] Step S514: Determine that the Intimate Payment channel poses a payment risk.

[0336] Among them, the relationship between user B and user A is...<u,v> And the set of users z who may have a marital relationship with user B. u The k1th user Extract the feature vectors of these three elements at the user level, such as the user's date of birth, asset information, and consumption preferences, and then concatenate these feature vectors to obtain... Furthermore, extract the relationship vector representing the user relationship between user B and user A, and extract the relationship vector representing the relationship between user B and user A. The relationship vectors of user relationships can be concatenated to obtain the following:

[0337] Similarly, for user B's n potential spouses, a set of feature vectors is obtained. and relation vector set For a user A, find m potential spouses to obtain a set of feature vectors. and relation vector set

[0338] Map the feature vectors from the two feature vector sets and the relation vectors from the two relation vector sets obtained above to a low-dimensional vector space, respectively. Then, concatenate the mapped feature vectors of user B with the relation vectors to obtain a concatenated vector set. And concatenate the feature vector mapped from user A with the relation vector to obtain a set of concatenated vectors.

[0339] like Figure 6 The model framework of the relationship prediction model shown adds a fully connected neural network to each concatenated vector in the two concatenated vector sets to perform vector transformation:

[0340] and

[0341] in, and The weight matrix and bias represent user B and the i-th user who may have a spousal relationship with user B, ultimately resulting in two sets of transformation vectors:

[0342]

[0343] Considering that different users who may have a marital relationship with User B contribute differently to the prediction of the current relationship's dissolution, they cannot be treated equally. Therefore, an attention mechanism is introduced to learn the attention weight of each user who may have a marital relationship with User B. Specifically, a two-layer fully connected neural network is used to implement the attention mechanism.

[0344] and

[0345] Among them, W 1 W 2 W 3 W 4 and b 1 b 2 b 3 b 4 These are parameters obtained in advance through model training;

[0346] and Merging can yield And calculate the attention weights according to the following formula:

[0347]

[0348] After calculating the attention weights, based on β u-＞v The overall normalized weights can be calculated:

[0349]

[0350] in,

[0351] Finally, the predicted score is calculated using a single-layer fully connected neural network.

[0352]

[0353] wherein, W and b represent weight matrix and bias respectively.

[0354] It should be noted that in the process of relationship detection and processing of the payment user and the associated user, the three processing methods of interaction type detection, interaction decay calculation and relationship dissolution prediction can be used simultaneously and sequentially for relationship detection and processing. In addition, in the specific execution process, any one or any two of the three processing methods of interaction type detection, interaction decay calculation and relationship dissolution prediction can be selected for relationship detection and processing according to the specific execution needs. The execution order of the three processing methods of interaction type detection, interaction decay calculation and relationship dissolution prediction in the relationship detection and processing process can also be adjusted accordingly, and the three processing methods after the order adjustment are used for relationship detection and processing. Similarly, the execution order of any two of the three processing methods of interaction type detection, interaction decay calculation and relationship dissolution prediction in the relationship detection and processing process can also be adjusted accordingly, and the two processing methods after the order adjustment are used for relationship detection and processing.

[0355] It should be further noted that the processing process of interaction type detection, interaction decay calculation and relationship dissolution prediction provided in the embodiment is only illustrative. The specific processing process and execution example of interaction type detection, interaction decay calculation and relationship dissolution prediction can refer to the corresponding content provided in steps S104 to S108 of the above method embodiment, and the embodiment will not be described here.

[0356] In step S1206, if the relationship detection fails, the protocol payment channel is marked and processed, and the payment channel set carrying the protocol payment channel is issued to the user terminal.

[0357] In the embodiment, by marking and processing the protocol payment channel, the payment user can perceive that the relationship between the associated user who opens the protocol payment channel and the payment user is abnormal or dissolved, thereby improving the payment success rate of the payment user. For the associated user who opens the protocol payment channel for the payment user, the probability of financial loss caused by the use of the protocol payment channel under the condition of change or dissolution of the user relationship is reduced, and the financial safety of the associated user is ensured.

[0358] In an optional implementation provided by the embodiment, the marking and processing of the protocol payment channel includes marking the protocol payment channel as an abnormal payment state, and / or degrading the payment priority of the protocol payment channel based on the abnormal payment state.

[0359] Step S1208, according to the payment processing request of the protocol payment channel sent by the user terminal, the payment confirmation of the protocol payment channel is performed for the associated user.

[0360] In specific implementation, on the basis of reducing the probability of causing capital loss caused by the use of the protocol payment channel in the case of changes in user relationship or termination, and ensuring the safety of the funds of the associated user, in order to further improve the payment success rate of the payment user, the embodiment processes the payment instruction of the payment user by the way of payment confirmation to the associated user. Specifically, in an optional implementation provided by the embodiment, the payment confirmation of the protocol payment channel for the associated user includes: sending a payment confirmation message to the associated user, and receiving a confirmation instruction of the payment confirmation message; based on the confirmation instruction, performing payment processing through the protocol payment channel.

[0361] For example, in the payment channel list of user A, the payment priority of the intimate payment payment channel with user B is adjusted from the first to the last. After adjustment, the payment channel list of user A is as shown in Figure 7 At the same time, the intimate payment payment channel displays an abnormal prompt "there is a payment risk, and this payment needs to be confirmed", and configures an access interface "apply for payment confirmation to TA" for applying to user B for confirmation of the current payment. If user A clicks the access interface "apply for payment confirmation to TA", an application payment page as shown in Figure 8 will be generated. After user A clicks the "confirm" button, a confirmation payment prompt message will be sent to user B. After user B triggers the confirmation payment prompt message, the payment confirmation page as shown in Figure 9 is entered.

[0362] The following takes the application of the payment detection processing method provided by the embodiment in the intimate payment scenario as an example, combined with Figure 11 , the payment detection processing method provided by the embodiment is further described, referring to Figure 11 , the payment detection processing method applied in the intimate payment scenario specifically includes the following steps.

[0363] Step S1106, according to the payment access request submitted by the user terminal of user A, the intimate payment channel for intimate payment opened by user B to user A is obtained.

[0364] Step S1108, based on the first interaction data of user A and user B, the interaction type detection is performed.

[0365] If the detection fails, step S1110 is performed.

[0366] If the detection passes, it is confirmed that the intimate payment channel does not have a payment risk.

[0367] Step S1110, interaction attenuation calculation is performed based on the second interaction data of user A and user B;

[0368] If the interaction attenuation value obtained by calculation is less than the attenuation threshold, step S1112 is performed;

[0369] If the interaction attenuation value obtained by calculation is greater than or equal to the attenuation threshold, step S1114 is performed;

[0370] Step S1112, a relationship prediction model is called to predict a relationship resolution value of user A and user B;

[0371] If the relationship resolution value exceeds the relationship resolution threshold, step S1114 is performed;

[0372] If the relationship resolution value does not exceed the relationship resolution threshold, it is confirmed that the intimate payment channel does not have payment risk.

[0373] Step S1114, if the intimate payment channel has payment risk, the intimate payment channel is marked.

[0374] Step S1116, the payment channel set containing the marked intimate payment channel is issued to the user terminal of user A.

[0375] Step S1124, a payment confirmation message is sent to user B according to the processing request for payment of the intimate payment channel.

[0376] After receiving the payment confirmation message, user B can enter a payment confirmation page by clicking the payment confirmation message, and if user A agrees to use the intimate payment, the user can submit a confirmation instruction to the server by triggering the payment confirmation control configured in the payment confirmation page.

[0377] Step S1126, according to the confirmation instruction, the payment processing of the order payment is performed through the intimate payment channel.

[0378] The payment detection processing device provided in the specification implements, for example:

[0379] In the above embodiment, a payment detection processing method is provided, and a payment detection processing device corresponding thereto is also provided, which will be described below with reference to the accompanying drawings.

[0380] Reference Figure 13 which shows a schematic diagram of a payment detection processing device provided in the embodiment.

[0381] Since the device embodiment corresponds to the method embodiment, the description is relatively simple, and the related parts can be referred to the above-mentioned corresponding description of the method embodiment. The device embodiments described below are only schematic.

[0382] The embodiment provides a payment detection processing device, which comprises:

[0383] The protocol payment channel obtaining module 1302 is configured to obtain a protocol payment channel opened by an associated user for a payment user;

[0384] The interaction type detection module 1304 is configured to perform interaction type detection based on first interaction data of the payment user and the associated user;

[0385] The interaction attenuation calculation module 1306 is configured to, if the detection fails, perform interaction attenuation calculation based on second interaction data of the payment user and the associated user;

[0386] The relationship release prediction module 1308 is configured to, if the calculated interaction attenuation value is less than an attenuation threshold value, invoke a relationship prediction model to perform relationship release prediction of the payment user and the associated user;

[0387] The marking processing module 1310 is configured to, if it is determined that the relationship is released based on the prediction result, perform marking processing on the protocol payment channel.

[0388] The second payment detection processing device provided in the specification is implemented as follows:

[0389] In the above embodiment, the second payment detection processing method is provided, and the second payment detection processing device corresponding to the method is also provided, which will be described below with reference to the accompanying drawings.

[0390] Reference is made to Figure 14 which shows a payment detection processing device provided in the embodiment.

[0391] Since the device embodiment corresponds to the method embodiment, the description is relatively simple, and the related parts can be seen in the above-mentioned corresponding description of the method embodiment. The device embodiment described below is only illustrative.

[0392] The embodiment provides a payment detection processing device, which runs on a user terminal of a payment user, and the device comprises:

[0393] The payment access request submission module 1402 is configured to submit a payment access request to a server based on a payment operation of the payment user;

[0394] The payment channel display module 1404 is configured to receive a payment channel set containing a protocol payment channel issued by the server and display the payment channel set; the protocol payment channel is marked after interaction type detection, interaction attenuation calculation and relationship release prediction of the payment user and an associated user opening the protocol payment channel are performed.

[0395] The payment processing request sending module 1406 is configured to send a payment processing request to the server to perform payment confirmation of the protocol payment channel for the associated user if payment instructions for the protocol payment channel are detected.

[0396] The third payment detection processing device provided in the specification implements the following, for example:

[0397] In the above embodiment, a third payment detection processing method is provided, and a third payment detection processing device corresponding thereto is also provided, which will be described below with reference to the accompanying drawings.

[0398] Referring to Figure 15 which shows a payment detection processing device provided in the embodiment.

[0399] Since the device embodiment corresponds to the method embodiment, it is described more simply, and the related part can be seen from the above-mentioned corresponding description of the method embodiment. The device embodiment described below is only illustrative.

[0400] The embodiment provides a payment detection processing device running on a server, and the device comprises:

[0401] The protocol payment channel acquisition module 1502 is configured to acquire a protocol payment channel opened by an associated user to a payment user according to a payment access request submitted by a user terminal of the payment user;

[0402] The relationship detection processing module 1504 is configured to perform relationship detection processing of the payment user and the associated user based on interaction data of the payment user and the associated user; the relationship detection processing comprises interaction type detection, interaction attenuation calculation and relationship removal prediction;

[0403] The marking processing module 1506 is configured to perform marking processing on the protocol payment channel if the relationship detection fails, and to issue a payment channel set carrying the protocol payment channel to the user terminal;

[0404] The payment confirmation module 1508 is configured to perform payment confirmation of the protocol payment channel for the associated user according to a payment processing request of the protocol payment channel sent by the user terminal.

[0405] The payment detection processing device provided in the specification implements the following, for example:

[0406] Corresponding to the above-mentioned payment detection processing method, based on the same technical concept, one or more embodiments of the specification also provide a payment detection processing device for executing the above-mentioned payment detection processing method,Figure 16 A structural schematic diagram of a payment detection processing device provided for one or more embodiments of the present specification.

[0407] The payment detection processing device provided by the embodiment comprises:

[0408] As shown in Figure 16 The payment detection processing device can have great differences due to different configurations or performances, and can include one or more processors 1601 and memories 1602, and one or more storage applications or data can be stored in the memories 1602. The memory 1602 can be temporary storage or persistent storage. The application stored in the memory 1602 can include one or more modules (not shown in the figure), and each module can include a series of computer executable instructions in the payment detection processing device. Further, the processor 1601 can be configured to communicate with the memory 1602 and execute a series of computer executable instructions in the memory 1602 on the payment detection processing device. The payment detection processing device can also include one or more power supplies 1603, one or more wired or wireless network interfaces 1604, one or more input / output interfaces 1605, one or more keyboards 1606, etc.

[0409] In a specific embodiment, the payment detection processing device includes a memory and one or more programs, wherein one or more programs are stored in the memory, and one or more programs can include one or more modules, and each module can include a series of computer executable instructions in the payment detection processing device, and the one or more processors are configured to execute the one or more programs include computer executable instructions for:

[0410] Obtaining an agreement payment channel opened by an associated user to a payment user;

[0411] Performing interaction type detection based on first interaction data of the payment user and the associated user;

[0412] If the detection fails, performing interaction decay calculation based on second interaction data of the payment user and the associated user;

[0413] In the case where the obtained interaction decay value is less than the decay threshold, calling a relationship prediction model to predict the relationship between the payment user and the associated user;

[0414] If it is determined that the relationship is removed based on the prediction result, performing marking processing on the agreement payment channel.

[0415] The second payment detection processing device provided by the present specification is as follows:

[0416] According to the second payment detection processing method described above, based on the same technical concept, one or more embodiments of the present specification further provide a second payment detection processing device for executing the payment detection processing method provided above, Figure 17 The structural schematic diagram of the second payment detection processing device provided by one or more embodiments of the present specification is shown.

[0417] The payment detection processing device provided by the embodiment includes:

[0418] As shown in Figure 17 The payment detection processing device can have great differences due to different configurations or performances, and can include one or more processors 1701 and memories 1702. The memory 1702 can store one or more storage applications or data. The memory 1702 can be temporary storage or persistent storage. The application stored in the memory 1702 can include one or more modules (not shown in the figure), and each module can include a series of computer executable instructions in the payment detection processing device. Further, the processor 1701 can be configured to communicate with the memory 1702 and execute a series of computer executable instructions in the memory 1702 on the payment detection processing device. The payment detection processing device can also include one or more power supplies 1703, one or more wired or wireless network interfaces 1704, one or more input / output interfaces 1705, one or more keyboards 1706, etc.

[0419] In a specific embodiment, the payment detection processing device includes a memory and one or more programs, wherein one or more programs are stored in the memory, and the one or more programs can include one or more modules, and each module can include a series of computer executable instructions in the payment detection processing device, and the one or more processors are configured to execute the one or more programs, which include the following computer executable instructions:

[0420] submitting a payment access request to a server based on a payment operation of a payment user;

[0421] receiving and displaying a payment channel set containing a protocol payment channel issued by the server; the protocol payment channel is marked after detecting the interaction type of the payment user and the associated user who opens the protocol payment channel, calculating the interaction attenuation, and predicting the relationship removal;

[0422] If a payment instruction for the protocol payment channel is detected, a payment processing request is sent to the server to perform payment confirmation of the protocol payment channel for the associated user.

[0423] The third payment detection processing device provided by the present specification implements, for example, the following:

[0424] Corresponding to the third payment detection processing method described above, based on the same technical concept, one or more embodiments of the present specification also provide a third payment detection processing device for executing the payment detection processing method provided above, Figure 18 The structural diagram of the third payment detection processing device provided by one or more embodiments of the present specification.

[0425] The payment detection processing device provided by the present embodiment includes:

[0426] As Figure 18 indicated, the payment detection processing device can have relatively large differences due to different configurations or performance, and can include one or more processors 1801 and memories 1802, and one or more storage applications or data can be stored in the memories 1802. Among them, the memory 1802 can be temporary storage or persistent storage. The application stored in the memory 1802 can include one or more modules (not shown in the figure), and each module can include a series of computer executable instructions in the payment detection processing device. Further, the processor 1801 can be configured to communicate with the memory 1802 and execute a series of computer executable instructions in the memory 1802 on the payment detection processing device. The payment detection processing device can also include one or more power supplies 1803, one or more wired or wireless network interfaces 1804, one or more input / output interfaces 1805, one or more keyboards 1806, etc.

[0427] In one specific embodiment, the payment detection processing device includes a memory and one or more programs, wherein one or more programs are stored in the memory, and one or more programs can include one or more modules, and each module can include a series of computer executable instructions in the payment detection processing device, and the one or more processors are configured to execute the one or more programs include computer executable instructions for:

[0428] According to the payment access request submitted by the user terminal of the payment user, the protocol payment channel opened by the associated user to the payment user is obtained;

[0429] based on the interaction data of the payment user and the associated user, performing relationship detection processing of the payment user and the associated user; the relationship detection processing includes interaction type detection, interaction attenuation calculation, and relationship release prediction;

[0430] If the relationship detection fails, marking the protocol payment channel, and issuing a payment channel set carrying the protocol payment channel to the user terminal;

[0431] According to the payment processing request of the protocol payment channel sent by the user terminal, performing payment confirmation of the protocol payment channel for the associated user.

[0432] The storage medium provided by the present specification implements, for example:

[0433] According to the above description, a payment detection processing method based on the same technical concept, one or more embodiments of the present specification also provide a storage medium.

[0434] The storage medium provided in this embodiment is used to store computer executable instructions, and the computer executable instructions realize the following processes when executed by a processor.

[0435] Obtain the protocol payment channel opened by the associated user to the payment user;

[0436] Based on the first interaction data of the payment user and the associated user, perform interaction type detection;

[0437] If the detection fails, based on the second interaction data of the payment user and the associated user, perform interaction attenuation calculation;

[0438] In the case where the obtained interaction attenuation value is less than the attenuation threshold, call the relationship prediction model to perform relationship release prediction of the payment user and the associated user;

[0439] If it is determined that the relationship is released based on the prediction result, marking the protocol payment channel.

[0440] It should be noted that the embodiments of the storage medium in the present specification and the embodiments of the payment detection processing method in the present specification are based on the same inventive concept, so the specific implementation of this embodiment can be referred to the implementation of the corresponding method described above, and the repeated parts will not be described again.

[0441] The second storage medium provided by the present specification implements, for example:

[0442] According to the second payment detection processing method described above, based on the same technical concept, one or more embodiments of the present specification also provide a second storage medium.

[0443] The storage medium provided by the embodiment is used for storing computer executable instructions, and the computer executable instructions realize the following process when executed by a processor.

[0444] A payment access request is submitted to a server based on a payment operation of the payment user.

[0445] A payment channel set containing a protocol payment channel issued by the server is received and displayed; the protocol payment channel is marked after interaction type detection, interaction attenuation calculation and relationship removal prediction are performed on the payment user and an associated user who opens the protocol payment channel.

[0446] If a payment instruction for the protocol payment channel is detected, a payment processing request is sent to the server to perform payment confirmation of the protocol payment channel for the associated user.

[0447] It should be noted that the embodiment of the second storage medium in the present specification is based on the same inventive concept as the embodiment of the second payment detection processing method in the present specification, and therefore the specific implementation of this embodiment can be referred to the implementation of the corresponding method described above, and the repeated parts will not be described again.

[0448] The third storage medium provided by the present specification is as follows:

[0449] According to the third payment detection processing method described above, based on the same technical concept, one or more embodiments of the present specification also provide a third storage medium.

[0450] The storage medium provided by the embodiment is used for storing computer executable instructions, and the computer executable instructions realize the following process when executed by a processor.

[0451] According to a payment access request submitted by a user terminal of a payment user, a protocol payment channel opened by an associated user to the payment user is acquired;

[0452] Based on the interaction data of the payment user and the associated user, relationship detection processing of the payment user and the associated user is performed; the relationship detection processing includes interaction type detection, interaction attenuation calculation and relationship removal prediction.

[0453] If the relationship detection fails, the protocol payment channel is marked, and a payment channel set carrying the protocol payment channel is issued to the user terminal;

[0454] According to the payment processing request of the protocol payment channel sent by the user terminal, payment confirmation of the protocol payment channel for the associated user is performed.

[0455] It should be noted that the embodiment of the third storage medium in the specification is based on the same inventive concept as the embodiment of the third payment detection processing method in the specification, and therefore the specific implementation of this embodiment can be referred to the implementation of the corresponding method as described above, and the repeated parts will not be described herein.

[0456] The above described particular embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the acts or steps recited in the claims can be performed in an order other than the order in which they are recited and still achieve desirable results. Additionally, the process depicted in the figures does not necessarily require the particular order shown, or sequential order, to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous or necessary.

[0457] In the 1930s, it was clear to distinguish whether an improvement in a technology was in hardware (e.g., improvement in circuit structure of diodes, transistors, switches, etc.) or in software (e.g., improvement in method flow). However, as technology has evolved, many improvements in method flow today can be considered as direct improvements in hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented by a hardware entity module. For example, a programmable logic device (PLD) (e.g., a field programmable gate array (FPGA)) is an integrated circuit whose logic function is determined by user programming of the device. A digital system is "integrated" on a PLD by the designer programming it, rather than by asking a chip manufacturer to design and fabricate a custom integrated circuit chip. Moreover, instead of manually fabricating an integrated circuit chip, this programming is now mostly implemented by "logic compiler" software, which is similar to software compilers used in program development, and the original code to be compiled is written in a specific programming language, called a hardware description language (HDL), of which there are many, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc., the most commonly used being VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. It should be clear to those skilled in the art that, by simply logically programming a method flow in one of the above hardware description languages and programming it into an integrated circuit, a hardware circuit implementing the logical method flow can be easily obtained.

[0458] The controller can be implemented in any suitable way, e.g. the controller can take the form of a microprocessor or processor and a computer readable medium storing computer readable program code, e.g. software or firmware, executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller and an embedded microcontroller, examples of controllers include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91 SAM, Microchip PIC18F26K20 and Silicone Labs C8051F320, the memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also know that in addition to being implemented in pure computer readable program code form, the controller can perfectly well be implemented by means of logic programmed into logic gates, switches, application specific integrated circuits, programmable logic controllers and embedded microcontrollers, etc. to perform the same functions. The controller can thus be considered a hardware component, and the means comprised therein for performing various functions can be considered structures within the hardware component. Alternatively, or even additionally, the means for performing various functions can be considered both software modules implementing the method and structures within the hardware component.

[0459] The systems, apparatuses, modules or units illustrated by the above embodiments can be implemented by computer chips or entities, or products with certain functions. A typical implementation device is a computer. Specifically, the computer can be a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0460] For the sake of description, the above apparatuses are described in various units with functions respectively. Of course, the functions of the units can be implemented in one or more software and / or hardware in implementing the embodiments of the present specification.

[0461] Those skilled in the art will appreciate that one or more embodiments of the present specification can be provided as a method, a system or a computer program product. Therefore, one or more embodiments of the present specification can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present specification can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage etc.) containing computer-usable program code.

[0462] The specification is presented with reference to flow diagrams and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the specification. It will be understood that each block of the flow diagrams and / or block diagrams, and combinations of blocks in the flow diagrams and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing element or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more flow or multiple flows and / or blocks Figure 1 one or more flow or multiple flows and / or blocks

[0463] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more flow or multiple flows and / or blocks Figure 1 one or more flow or multiple flows and / or blocks

[0464] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more flow or multiple flows and / or blocks Figure 1 one or more flow or multiple flows and / or blocks

[0465] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0466] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) and / or cache memory, non-volatile memory, such as read-only memory (ROM), EPROM, and / or flash memory, etc. The memory is an example of computer readable media.

[0467] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0468] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not only include those elements, but also other elements not explicitly listed, or other elements inherent to such a process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0469] One or more embodiments of the specification can be described in the general context of computer-executable instructions being executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of the specification can also be practiced in a distributed computing environment, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0470] Various embodiments in the specification are described in a progressive manner, and the same or similar parts between various embodiments can be referred to each other, and each embodiment focuses on the difference from other embodiments. In particular, for system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0471] The above merely provides the example of the present document and is not intended to limit the present document. For those skilled in the art, the present document can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present document shall be included in the scope of claims of the present document.

Claims

1. A payment detection processing method, comprising: obtaining a protocol payment channel opened by an associated user to a payment user, the protocol payment channel binding a payment account of the associated user; detecting whether an interaction type carried by an interaction record of the payment user and the associated user is a negative interaction type; if not, detecting whether a generation time length of the interaction record is greater than a preset time threshold, if greater, performing interaction attenuation calculation based on interaction data of the payment user and the associated user; in the case that the obtained interaction attenuation value is less than the attenuation threshold, calling a relationship prediction model to predict the relationship between the payment user and the associated user, the relationship prediction model including: predicting the probability of the user relationship between the payment user and the associated user being removed by detecting the user relationship with uniqueness or exclusivity; if the relationship is removed based on the prediction result, marking the protocol payment channel.

2. The payment detection processing method of claim 1, wherein the interaction data comprises graph structure interaction data stored in a graph database; wherein the data nodes in the graph structure interaction data correspond to the user identifier of the payment user and the user identifier of the associated user; the data connection in the graph structure interaction data corresponds to a data sequence composed of the user identifier of the payment user, the user identifier of the associated user, the interaction type and / or the time information.

3. The payment detection processing method of claim 1, wherein the interaction attenuation calculation based on the interaction data of the payment user and the associated user comprises: calculating the cumulative interaction frequency of the payment user and the associated user based on the total number of interactions contained in the interaction data; calculating a time attenuation score based on the average activity and interaction interval length of the payment user and the associated user; calculating the interaction attenuation value according to the cumulative interaction frequency and the time attenuation score.

4. The payment detection processing method of claim 1, wherein if the detection fails, after the interaction attenuation calculation step based on the interaction data of the payment user and the associated user is executed, further comprising: if the obtained interaction attenuation value is greater than or equal to the attenuation threshold, marking the protocol payment channel.

5. The payment detection processing method of claim 1, wherein the calling the relationship prediction model to predict the relationship between the payment user and the associated user comprises: obtaining first user data of a first associated user set of the payment user and second user data of a second associated user set of the associated user; inputting the first user data and the second user data into the relationship prediction model to predict the relationship removal value of the payment user and the associated user as the prediction result.

6. The payment detection processing method of claim 5, wherein the relationship prediction model comprises a first neural network, a second neural network and a third neural network; correspondingly, the predicting the relationship removal value of the payment user and the associated user comprises: inputting the first spliced vector in the vector set of the payment user and the second spliced vector in the vector set of the associated user into corresponding first neural networks respectively for vector transformation to obtain transformed vectors; performing attention weight calculation on the transformed vectors through the second neural network, and calculating normalization weights according to the calculated attention weights; calculating weighted vectors according to the normalization weights and the transformed vectors, and inputting the weighted vectors into the third neural network for calculation to obtain the relationship resolution value.

7. The payment detection processing method according to claim 6, before the sub-step of inputting the first spliced vector in the vector set of the payment user and the second spliced vector in the vector set of the associated user into corresponding first neural networks respectively for vector transformation to obtain transformed vectors, further comprising: performing vector mapping on the initial feature vector and the initial relationship vector of the payment user to obtain a first feature mapping vector and a first relationship mapping vector; performing vector splicing on the first feature mapping vector and the first relationship mapping vector to obtain the first spliced vector; and / or, performing vector mapping on the initial feature vector and the initial relationship vector of the associated user to obtain a second feature mapping vector and a second relationship mapping vector; performing vector splicing on the second feature mapping vector and the second relationship mapping vector to obtain the second spliced vector.

8. The payment detection processing method according to claim 7, wherein the initial feature vector of the payment user is obtained in the following manner: obtaining a first feature vector of the payment user, a second feature vector of the associated user, and a third feature vector of a candidate associated user of the payment user; performing vector splicing on the first feature vector, the second feature vector, and the third feature vector to obtain the initial feature vector of the payment user.

9. The payment detection processing method according to claim 7, wherein the initial relationship vector of the payment user is obtained in the following manner: obtaining a first interaction relationship vector between the payment user and the associated user, and a second interaction relationship vector between the payment user and a candidate associated user; performing splicing on the first interaction relationship vector and the second interaction relationship vector to obtain the initial relationship vector of the payment user.

10. The payment detection processing method of claim 6, the attention weight comprising: a first attention weight of the payment user and a candidate associated user of the payment user, and a second attention weight of the associated user and a candidate associated user of the associated user; correspondingly, the calculation of the normalization weights according to the calculated attention weights comprises: calculating a weight sum of the first attention weight and the second attention weight; calculating a ratio of the first attention weight to the weight sum as the normalization weight.

11. The payment detection processing method according to claim 1, after the operation of invoking the relationship prediction model to perform the relationship resolution prediction of the payment user and the associated user, further comprising: if it is determined based on the prediction result that the relationship between the payment user and the associated user has not been resolved, confirming that the payment state of the agreement payment channel is normal.

12. The payment detection processing method of claim 1, wherein the marking processing of the protocol payment channel comprises: marking the protocol payment channel as an abnormal payment state; and degrading the payment priority of the protocol payment channel based on the abnormal payment state.

13. The payment detection processing method of claim 1, wherein after the marking processing of the protocol payment channel is performed based on the prediction result, the method further comprises: sending a payment confirmation message to the associated user according to a payment instruction of the protocol payment channel; and performing payment processing based on the protocol payment channel if a confirmation instruction of the payment confirmation message is detected.

14. A payment detection processing method applied to a user terminal of a payment user, the method comprising: submitting a payment access request to a server based on a payment operation of the payment user; receiving and displaying a payment channel set including a protocol payment channel issued by the server; the protocol payment channel binds a payment account of an associated user; the protocol payment channel is marked after interaction type detection, interaction attenuation calculation and relationship release prediction of the payment user and the associated user who opens the protocol payment channel are performed; the relationship release prediction comprises predicting a probability of user relationship release between the payment user and the associated user by detecting a user relationship with uniqueness or exclusivity; if a payment instruction for the protocol payment channel is detected, a payment processing request is sent to the server to perform payment confirmation of the protocol payment channel for the associated user.

15. The payment detection processing method of claim 14, wherein the marking processing comprises: marking the protocol payment channel as an abnormal payment state; and degrading the payment priority of the protocol payment channel based on the abnormal payment state. Correspondingly, the payment confirmation of the protocol payment channel for the associated user comprises: sending a payment confirmation message to the associated user and receiving a confirmation instruction of the payment confirmation message; and performing payment processing based on the confirmation instruction through the protocol payment channel.

16. The payment detection processing method of claim 14, wherein the interaction type detection is implemented in the following manner: detecting whether an interaction type carried by first interaction data of the payment user and the associated user is a negative interaction type; if the interaction type is a negative interaction type, marking the protocol payment channel; and if the interaction type is a positive interaction type, detecting whether a generation time length of the first interaction data is greater than a preset time threshold, and if yes, performing the interaction attenuation calculation.

17. The payment detection processing method of claim 14, wherein the interaction attenuation calculation is implemented in the following manner: calculating a cumulative interaction frequency of the payment user and the associated user based on a total interaction times included in second interaction data of the payment user and the associated user; and calculating a time attenuation score based on an average activity and an interaction interval time length of the payment user and the associated user included in the second interaction data. ​ ​ An interaction decay value is calculated according to the accumulated interaction frequency and the time decay score, and the relationship resolution prediction is performed if the interaction decay value is less than a decay threshold.

18. The payment detection processing method of claim 14, wherein the relationship resolution prediction is implemented in the following manner: a first spliced vector in the vector set of the payment user and a second spliced vector in the vector set of the associated user are respectively input into corresponding first neural networks for vector transformation to obtain transformed vectors; attention weight calculation of the transformed vectors is performed by a second neural network, and a normalization weight is calculated according to the calculated attention weight; a weighted vector is calculated according to the normalization weight and a feature vector, the weighted vector is input into a third neural network for calculation, and the marking processing is performed if the obtained relationship resolution value is greater than a preset threshold.

19. A payment detection processing method applied to a server, the method comprising: obtaining, according to a payment access request submitted by a user terminal of a payment user, a protocol payment channel opened by an associated user to the payment user, the protocol payment channel being bound to a payment account of the associated user; performing relationship detection processing of the payment user and the associated user based on interaction data of the payment user and the associated user; the relationship detection processing comprises interaction type detection, interaction decay calculation and relationship resolution prediction, and the relationship resolution prediction comprises predicting a probability of resolution of a user relationship between the payment user and the associated user by detecting a user relationship with uniqueness or exclusivity; if the relationship detection fails, performing marking processing on the protocol payment channel, and issuing a payment channel set carrying the protocol payment channel to the user terminal; performing payment confirmation of the protocol payment channel for the associated user according to a payment processing request of the protocol payment channel sent by the user terminal.

20. A payment detection processing device, comprising: a protocol payment channel obtaining module configured to obtain a protocol payment channel opened by an associated user to a payment user, the protocol payment channel being bound to a payment account of the associated user; an interaction type detection module configured to detect whether an interaction type carried by an interaction record of the payment user and the associated user is a negative interaction type, and if not, detect whether a generation time length of the interaction record is greater than a preset time threshold, and if so, run an interaction decay calculation module; the interaction decay calculation module is configured to perform interaction decay calculation based on interaction data of the payment user and the associated user; a relationship resolution prediction module configured to call a relationship prediction model to perform relationship resolution prediction of the payment user and the associated user if an interaction decay value calculated is less than a decay threshold, the relationship resolution prediction comprising predicting a probability of resolution of a user relationship between the payment user and the associated user by detecting a user relationship with uniqueness or exclusivity; a marking processing module configured to perform marking processing on the protocol payment channel if it is determined that the relationship is resolved based on the prediction result. 21.A payment detection processing apparatus, operating on a user terminal of a payment user, comprising: a payment access request submitting module configured to submit a payment access request to a server based on a payment operation of the payment user; a payment channel display module configured to receive and display a payment channel set containing a protocol payment channel issued by the server; the protocol payment channel is bound to a payment account of an associated user; the protocol payment channel is marked after detecting interaction type of the payment user and the associated user, calculating interaction attenuation, and predicting relationship removal; the relationship removal prediction includes predicting probability of user relationship removal between the payment user and the associated user by detecting user relationship with uniqueness or exclusivity; a payment processing request sending module configured to send a payment processing request to the server to confirm payment of the protocol payment channel for the associated user if a payment instruction for the protocol payment channel is detected. 22.A payment detection processing apparatus, operating on a server, comprising: a protocol payment channel obtaining module configured to obtain a protocol payment channel opened by an associated user to a payment user according to a payment access request submitted by a user terminal of the payment user; the protocol payment channel is bound to a payment account of the associated user; a relationship detection processing module configured to perform relationship detection processing between the payment user and the associated user based on interaction data of the payment user and the associated user; the relationship detection processing includes interaction type detection, interaction attenuation calculation, and relationship removal prediction; the relationship removal prediction includes predicting probability of user relationship removal between the payment user and the associated user by detecting user relationship with uniqueness or exclusivity; a marking processing module configured to mark the protocol payment channel if relationship detection fails, and issue a payment channel set carrying the protocol payment channel to the user terminal; a payment confirmation module configured to confirm payment of the protocol payment channel for the associated user according to a payment processing request of the protocol payment channel sent by the user terminal. 23.A payment detection processing device, comprising: a processor; and a memory configured to store computer executable instructions which, when executed, cause the processor to: obtain a protocol payment channel opened by an associated user to a payment user; the protocol payment channel is bound to a payment account of the associated user; detect whether an interaction type carried by an interaction record of the payment user and the associated user is a negative interaction type; if not, detect whether a generation duration of the interaction record is greater than a preset time threshold; if yes, perform interaction attenuation calculation based on interaction data of the payment user and the associated user; In a case where the obtained interaction attenuation value is less than the attenuation threshold, a relationship prediction model is called to perform relationship removal prediction of the payment user and the associated user, and the relationship removal prediction includes: predicting a probability of user relationship removal of the payment user and the associated user by detecting a user relationship with uniqueness or exclusivity. If it is determined that the relationship is removed based on the prediction result, the protocol payment channel is marked.

24. A storage medium for storing computer executable instructions, the computer executable instructions, when executed by a processor, implement the following processes: obtain a protocol payment channel opened by an associated user to a payment user, the protocol payment channel binding a payment account of the associated user; detect whether an interaction type carried by an interaction record of the payment user and the associated user is a negative interaction type; if not, detect whether a generation time length of the interaction record is greater than a preset time threshold, if greater, perform interaction attenuation calculation based on interaction data of the payment user and the associated user; in a case where an obtained interaction attenuation value is less than an attenuation threshold, a relationship prediction model is called to perform relationship removal prediction of the payment user and the associated user, and the relationship removal prediction includes: predicting a probability of user relationship removal of the payment user and the associated user by detecting a user relationship with uniqueness or exclusivity; if it is determined that the relationship is removed based on the prediction result, the protocol payment channel is marked.

Images