Payment processing method and device

Abstract

One or more embodiments of the invention disclose a payment processing method and device. The method comprises the following steps: acquiring first order information of a to-be-processed order; inputting the first order information into a pre-trained multi-target prediction model, and outputting prediction values of the to-be-processed order in different staging modes of different prediction targets; the prediction value is used for representing a prediction result of payment of the to-be-processed order by using a staging mode of the prediction target; the prediction target comprises at least one of installment rights and interests, installment channels and installment periods; the staging mode comprises at least one of sub-staging rights and interests, sub-staging channels and sub-staging periods; determining target installment strategy information corresponding to the to-be-processed order according to the predicted value, and sending the target installment strategy information to the cashier platform; the target installment strategy information is composed of a target installment mode in the prediction target, and is used for the cashier platform to determine recommended payment mode information according to the first order information and the target installment strategy information.

Description

Technical Field

[0001] This specification relates to the field of data processing technology, and in particular to a payment processing method and device. Background Art

[0002] With the rapid development of e-commerce and financial technology, the variety of products traded is growing, and the amounts involved are also increasing. Installment payments can, to a certain extent, increase users' willingness to participate in transactions. However, for a pending order, current checkout platforms typically display various payment methods in a single recommended order. Due to the lack of information on installment strategies for the pending order, users are unable to understand the advantages of installment payments, resulting in a low likelihood of choosing them. Summary of the Invention

[0003] On the one hand, one or more embodiments of the present specification provide a payment processing method, which is applied to an installment strategy decision system, the method comprising: obtaining first order information of a pending order, inputting the first order information into a pre-trained multi-objective prediction model, and outputting predicted values ​​of the pending order under different installment methods with different prediction targets. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. The prediction target includes at least one of installment rights, installment channels, and installment period numbers. The installment method includes at least one of sub-installment rights, sub-installment channels, and sub-installment period numbers. The target installment strategy information corresponding to the pending order is determined based on the predicted value, and the target installment strategy information is sent to a cash register platform. The target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

[0004] On the other hand, one or more embodiments of the present specification provide a payment processing method, which is applied to a cash register platform, and the method includes: obtaining the first order information of a pending order, and receiving the target installment strategy information corresponding to the pending order sent by the installment strategy decision system. The target installment strategy information is obtained in the following manner: inputting the first order information of the pending order into a pre-trained multi-objective prediction model, outputting the predicted value of the pending order under different installment methods with different prediction targets, and determining the target installment strategy information corresponding to the pending order based on the predicted value. Based on the first order information, determine a plurality of candidate payment methods corresponding to the pending order, determine the recommended payment method information based on the plurality of candidate payment methods and the target installment strategy information, and display the recommended payment method information.

[0005] On another aspect, one or more embodiments of the present specification provide a payment processing device, which is applied to an installment strategy decision system, and the device includes: a first acquisition module, which is used to obtain the first order information of a pending order. A model prediction module, which is used to input the first order information into a pre-trained multi-objective prediction model, and output the predicted value of the pending order under different installment methods with different prediction targets. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. The prediction target includes at least one of installment rights, installment channels, and installment period numbers. The installment method includes at least one of sub-installment rights, sub-installment channels, and sub-installment period numbers. A determination and sending module, which is used to determine the target installment strategy information corresponding to the pending order based on the predicted value, and send the target installment strategy information to a cash register platform, wherein the target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

[0006] On another aspect, one or more embodiments of the present specification provide a payment processing device, comprising: a processor; and a memory arranged to store computer-executable instructions, wherein when the executable instructions are executed, the processor is enabled to: obtain first order information of a pending order, input the first order information into a pre-trained multi-objective prediction model, and output predicted values ​​of the pending order under different installment methods with different prediction targets. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. The prediction target includes at least one of installment benefits, installment channels, and installment period numbers. The installment method includes at least one of sub-installment benefits, sub-installment channels, and sub-installment period numbers. The target installment strategy information corresponding to the pending order is determined based on the predicted value, and the target installment strategy information is sent to a cash register platform. The target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

[0007] On the other hand, an embodiment of the present specification provides a storage medium for storing computer-executable instructions, which implement the following process when executed by a processor: obtaining the first order information of a pending order, inputting the first order information into a pre-trained multi-objective prediction model, and outputting the predicted value of the pending order under different installment methods with different prediction targets. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. The prediction target includes at least one of installment rights, installment channels, and installment period numbers. The installment method includes at least one of sub-installment rights, sub-installment channels, and sub-installment period numbers. The target installment strategy information corresponding to the pending order is determined based on the predicted value, and the target installment strategy information is sent to the cash register platform. The target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate one or more embodiments of this specification or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in one or more embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0009] Figure 1 is a schematic block diagram of a payment processing system according to an embodiment of this specification; Figure 2 is a schematic flow chart of a payment processing method according to an embodiment of this specification; Figure 3 This is a schematic diagram of a specific implementation process of obtaining a sample prediction value according to an embodiment of this specification; Figure 4 is a schematic flow chart of a payment processing method according to another embodiment of this specification; Figure 5 is a schematic swim lane diagram of a payment processing method according to an embodiment of this specification; Figure 6 is a schematic block diagram of a payment processing device according to an embodiment of this specification; Figure 7 is a schematic block diagram of a payment processing device according to another embodiment of the present specification; Figure 8 It is a structural diagram of a payment processing device according to an embodiment of this specification. DETAILED DESCRIPTION

[0010] One or more embodiments of this specification provide a payment processing method and apparatus to solve the current problem of being unable to make targeted installment strategy decisions for orders.

[0011] In order to help those skilled in the art better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of one or more embodiments of this specification.

[0012] As the variety of products traded grows and the amounts involved increase, installment payments can, to a certain extent, increase users' willingness to participate in transactions. However, for a pending order, current checkout platforms typically display information about various payment methods in a single recommended order. Due to a lack of information about installment strategies specific to the pending order, users are unable to understand the advantages of installment payments, resulting in a low likelihood of choosing them. Based on this, the embodiments of this specification provide a payment processing method and apparatus.

[0013] Figure 1 is a schematic block diagram of a payment processing system according to an embodiment of this specification, such as Figure 1 As shown, the payment processing system includes an installment strategy decision system 110 and a cash register platform 120. The installment strategy decision system 110 is used to obtain the first order information of the pending order, input the first order information into a pre-trained multi-objective prediction model, and output the predicted value of the pending order under different installment methods with different prediction targets. The predicted value is used to characterize the predicted result of payment for the pending order using the installment method of the prediction target. The prediction target includes at least one of the installment rights, installment channels, and installment periods. The installment method includes at least one of the sub-installment rights, sub-installment channels, and sub-installment periods. Thus, the target installment strategy information corresponding to the pending order is determined based on the predicted value, and the target installment strategy information is sent to the cash register platform 120. The target installment strategy information is composed of the target installment method in the prediction target.

[0014] The cash register platform 120 is used to obtain the first order information of the pending order, and to receive the target installment strategy information corresponding to the pending order sent by the installment strategy decision system 110, so as to determine the multiple candidate payment methods corresponding to the pending order based on the first order information, and determine the recommended payment method information based on the multiple candidate payment methods and the target installment strategy information, and then display the recommended payment method information.

[0015] Optionally, the cash register platform 120 is further configured to feed back first billing information of the processed order to the installment strategy decision system 110. The first billing information includes at least one of the following: first order identification information, installment benefits used for the processed order, installment channels used for the processed order, and the number of installments used for the processed order. The installment strategy decision system 110 is further configured to determine the target installment strategy information corresponding to the processed order based on the first order identification information, and to determine a first conversion result of the target installment strategy information corresponding to the processed order based on the first billing information and the target installment strategy information corresponding to the processed order, thereby determining the accuracy of the installment strategy decision system's decision on the installment method for the processed order based on the first conversion result and a preset effect evaluation strategy.

[0016] The following describes in detail the operations performed by the installment strategy decision system 110 and the cashier platform 120 in the payment processing system during the payment processing process. Figure 2 This is a schematic flow chart of a payment processing method according to an embodiment of this specification. In this embodiment, the payment processing method is applied to Figure 1 The staging strategy decision system 110 shown in FIG. Figure 2 As shown, the method may include: S202, obtaining the first order information of the order to be processed.

[0017] Among them, the pending orders are orders that support installment payments. The first order information may include the order amount, order creation time, order identification information (such as user identification information, product identification information, merchant identification information, order number, etc.), transaction type, supported sub-installment rights, supported sub-installment channels, and at least one of the supported sub-installment periods. The first order information of the pending orders can be sent by the downstream traffic pool to the installment strategy decision system. It can be understood that the orders in the downstream traffic pool include not only pending orders that support installment payments, but also orders that do not support installment payments. For orders that do not support installment payments, the downstream traffic pool will send them directly to the cash register platform for processing. For pending orders that support installment payments, the downstream traffic pool will send them to the installment strategy decision system and the cash register platform for processing respectively.

[0018] In the application, user identification information can be user name, user ID (Identity document, account number), etc., product identification information can be product name, product ID, etc., merchant identification information can be merchant name, merchant ID, etc., and transaction type can be online shopping, insurance, merchant code shopping, personal code shopping, etc.

[0019] Optionally, the user may initiate payment to trigger the installment strategy decision system to obtain the first order information of the pending order. Based on this, the step is performed as follows: in response to the user's request to make payment for the pending order, obtain the first order information of the pending order.

[0020] S204: Input the first order information into a pre-trained multi-objective prediction model, and output prediction values ​​of the pending order under different prediction objectives and different installment methods.

[0021] The predicted value represents the predicted result of payment for pending orders using the installment method with the predicted target. The predicted target may include at least one of installment benefits, installment channels, and the number of installments. The installment method may include at least one of sub-installment benefits, sub-installment channels, and the number of sub-installments.

[0022] In the application, sub-installment benefits may include interest-free, discounts, coupons, red envelopes, etc., sub-installment channels may include credit cards, credit purchases, Huabei, etc., and sub-installment periods may include 3 periods, 6 periods, 9 periods, 12 periods, etc. It is understood that installment methods include but are not limited to the above-mentioned ones.

[0023] Optionally, the multi-objective prediction model may be a CVR (Conversion Rate) model. The multi-objective prediction model may be constructed based on an MMoE (Multi-gate Mixture-of-Experts) model.

[0024] S206: Determine target installment strategy information corresponding to the pending order based on the predicted value, and send the target installment strategy information to the cash register platform.

[0025] The target installment strategy information is composed of the target installment method in the predicted target, and is used by the cashier platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

[0026] Adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the installment strategy decision system, the first order information is input into the pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output. The prediction target includes at least one of the installment rights, installment channels, and installment period numbers, and the installment method includes at least one of the sub-installment rights, sub-installment channels, and sub-installment period numbers. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. Thus, the target installment strategy information corresponding to the pending order is determined according to the predicted value, and the target installment strategy information is sent to the cash register platform. The target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information. As can be seen, this technical solution can utilize a multi-objective prediction model to specifically predict the user's payment results for the pending order using various installment methods for each prediction target based on the first order information of the pending order, thereby outputting the target installment strategy information determined based on the prediction results to the cash register platform, achieving targeted, accurate, and efficient installment strategy decisions for the pending order. This helps the cash register platform prioritize the installment payment method and corresponding installment strategy information in limited exposure locations, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments. In addition, since the cash register platform does not need to determine the installment strategy, the decision-making complexity of the cash register platform is reduced.

[0027] In one embodiment, a multi-objective prediction model can be trained according to the following steps A1 and A2.

[0028] Step A1: Obtain sample order information and sample billing information of multiple sample orders.

[0029] The sample order information may include at least one of the following: the sample order amount, the sample order creation time, the sample order identification information, the sample transaction type, the supported sub-installment benefits, the supported sub-installment channels, and the supported number of sub-installments. The sample billing information may include the installment benefits used in the sample order, the installment channels used in the sample order, and the number of installments used in the sample order.

[0030] In the application, the sample order identification information includes but is not limited to user identification information, product identification information, merchant identification information and order number.

[0031] In step A2, the sample order information and the sample bill information are input into the multi-objective prediction model to be trained for iterative training to obtain a trained multi-objective prediction model.

[0032] In this embodiment, by obtaining sample order information and sample billing information of multiple sample orders, the sample order information and sample billing information are input into the multi-objective prediction model to be trained for iterative training, and the trained multi-objective prediction model is obtained, which provides a model basis for making targeted, accurate and efficient installment strategy decisions for the orders to be processed.

[0033] In one embodiment, sample order information and sample bill information are input into the multi-objective prediction model to be trained for iterative training to obtain a trained multi-objective prediction model (i.e., step A2), which can be executed as follows: steps A21 to A23: Step A21: In the multi-objective prediction model to be trained, based on the sample order information, a sample prediction value of payment for the sample order using each installment method of each prediction objective is predicted.

[0034] For example, when the prediction targets of the multi-objective prediction model to be trained include installment benefits, installment channels, and installment periods, and the installment methods include sub-installment benefits, sub-installment channels, and sub-installment periods, if the sub-installment benefits include interest-free, discounts, and red envelopes, the sub-installment channels include credit cards and Huabei, and the sub-installment periods include 3, 6, 12, and 24 periods, then, in the multi-objective prediction model to be trained, based on the sample order information, sample prediction values ​​for paying the sample order using the three installment benefits of interest-free, discounts, and red envelopes can be predicted respectively, and sample prediction values ​​for paying the sample order using the two installment channels of credit cards and Huabei can be predicted respectively, and sample prediction values ​​for paying the sample order using the four installment periods of 3, 6, 12, and 24 periods can be predicted respectively. That is, sample prediction values ​​for paying the sample order using nine installment methods with three prediction targets can be obtained.

[0035] Step A22, for each sample order, based on the sample prediction value of the sample order in each installment mode, as well as the sample bill information and the loss function corresponding to the multi-objective prediction model, determine the loss value of predicting the sample order using the multi-objective prediction model.

[0036] Among them, the loss function of the multi-objective prediction model can be an applicable and reasonable loss function, and this embodiment does not limit this.

[0037] In practice, for each sample order, step A22 may be performed as follows: based on the sample billing information indicating the installment benefits, installment channels, and number of installments used for the sample order, the sample installment strategy information corresponding to the sample order is determined. Furthermore, based on the sample predicted values ​​for each installment method output from step A21, the predicted installment strategy information corresponding to the sample order is determined. Consequently, the sample installment strategy information and the predicted installment strategy information are substituted into the loss function of the multi-objective prediction model to obtain the loss value predicted for the sample order using the multi-objective prediction model.

[0038] Among them, according to the sample prediction value of the sample order under each installment method output in the above step A21, the prediction installment strategy information corresponding to the sample order is determined, which can be executed as follows: determining the prediction installment method corresponding to the maximum sample prediction value of the sample order under each prediction target, and forming the prediction installment strategy information based on the prediction installment method under each prediction target.

[0039] Continuing with the example in step A21, if the sample prediction value of the sample order under interest-free is 10%, the sample prediction value under discount is 15%, the sample prediction value under red envelope is 25%, the sample prediction value under credit card is 15%, the sample prediction value under Huabei is 35%, the sample prediction value under 3 installments is 5%, the sample prediction value under 6 installments is 15%, the sample prediction value under 12 installments is 20%, and the sample prediction value under 24 installments is 10%, then it can be determined that the predicted installment methods corresponding to the maximum sample prediction value of the sample order under each prediction target include red envelopes, Huabei and 12 installments, that is, the predicted installment strategy information is red envelopes, Huabei and 12 installments.

[0040] In step A23, the model parameters are iteratively adjusted according to the loss value until the iteration termination condition is met, and the training is stopped to obtain the trained multi-objective prediction model.

[0041] Optionally, the iteration termination condition may be that the loss value converges or reaches a preset number of iterations.

[0042] In this embodiment, a multi-objective prediction model is trained to provide a foundation for accurate and efficient staging strategies for targeted orders. Compared to multiple modeling approaches, this approach is not only more convenient to implement but also takes into account the differences and connections between different prediction objectives, resulting in better prediction results.

[0043] In one embodiment, the multi-objective prediction model may include an embedding layer, an expert network layer, a gating network layer, and neural network layers corresponding to different prediction objectives.

[0044] The embedding layer is used to embed the input discrete and continuous variables separately and then concatenate them together, undergoing several full concatenations to form a fixed-dimensional vector x. Specifically, the embedding layer discretizes continuous features into bins based on their dimensions and then performs feature embedding on them together with sparse features. The embedding results of features of different dimensions are then concatenated to map high-dimensional sparse features to low-dimensional dense features. Embedding is the process of mapping high-dimensional data (such as text) into a low-dimensional space.

[0045] The expert network layer can include m expert networks, each of which is fully connected with the same size. x passes through m expert networks respectively, and m expert vectors can be obtained.

[0046] The gating network layer can include multiple gating networks. In applications, a gating network can be learned for each phase of each prediction target. The gating network takes x as input and outputs a probability vector after softmax. This probability vector is then used to perform a weighted average of the m expert vectors to capture diverse information. Softmax is used to compress a real number vector into a probability distribution vector, such that each element is between (0, 1) and the sum of all elements is 1.

[0047] The neural network layer corresponding to each prediction target can be a multi-layered tower task network, used to predict the sample prediction value for a sample order payment in a specific installment method for that prediction target. The weighted average result serves as the input to the corresponding tower task network. After a full-connection mapping, it is converted to a one-dimensional vector for classification or regression, resulting in the sample prediction value.

[0048] For example, if the prediction target is the number of installments, and the installment methods include 3 installments, 6 installments, 12 installments, and 24 installments, then the multi-objective prediction model will include 4 tower task networks, each of which is used to output the sample prediction value for the corresponding installment method for the sample order payment.

[0049] In this embodiment, in the multi-objective prediction model to be trained, based on the sample order information, the sample prediction value of the sample order payment using each installment method of each prediction target is predicted (i.e., step A21), which can be performed as follows: Figure 3 As shown, the specific implementation process of obtaining the sample prediction value is shown: First, the input sample order information is subjected to dimensionality reduction processing through the embedding layer to obtain the sample order feature vector corresponding to the sample order information.

[0050] Optionally, the input sample order information may include the sub-installment channels supported by the sample order, the sub-installment rights supported, the user's basic portrait, etc. The user's basic portrait may include user identification information, the user's historical transaction behavior and financial behavior, the user's login behavior and access behavior in payment applications, etc.

[0051] Secondly, through the expert network layer, the input sample order feature vector is mapped to the first preset vector space to obtain the sample order expert vector corresponding to the sample order feature vector.

[0052] For example, Figure 3 As shown, the expert network layer includes 8 expert networks, ExpErt 1, ExpErt 2, ..., ExpErt 8. Then, after this step, 8 sample order expert vectors corresponding to the sample order feature vector can be obtained, namely, sample order expert vector 1, sample order expert vector 2, ..., sample order expert vector 8.

[0053] Again, through the gating network layer, the input sample order feature vector is mapped to the second preset vector space to obtain the mapping result, and the mapping result is used to perform weighted averaging on all sample order expert vectors to obtain the gated output result.

[0054] For example, Figure 3 As shown in the figure, the gated network layer includes four gate networks, Gate 1, Gate 2, Gate 3, and Gate 4. Then, after this step, four gated output results can be obtained, namely gated output result 1, gated output result 2, gated output result 3, and gated output result 4.

[0055] Afterwards, the input gated output result is mapped to a third preset vector space through a neural network layer to obtain a sample prediction value for payment of the sample order in an installment manner using a prediction target.

[0056] For example, Figure 3As shown, the neural network layer includes four tower task networks, Tower 1, Tower 2, Tower 3, and Tower 4. In this step, the gated output results will be used as the input of the corresponding tower task network. For example, gated output result 1 is the input of Tower 1. If Tower 1 is used to predict the sample prediction value of a 3-period installment, the probability of a 3-period installment can be obtained. Gated output result 2 is the input of Tower 2. If Tower 2 is used to predict the sample prediction value of a 6-period installment, the probability of a 6-period installment can be obtained. Gated output result 3 is the input of Tower 3. If Tower 3 is used to predict the sample prediction value of a 12-period installment, the probability of a 12-period installment can be obtained. Gated output result 4 is the input of Tower 4. If Tower 4 is used to predict the sample prediction value of a 24-period installment, the probability of a 24-period installment can be obtained.

[0057] In one embodiment, the decision accuracy of the staging strategy decision system may be determined according to the following steps B1 to B4.

[0058] Step B1: Receive first billing information of the processed order fed back by the cash register platform.

[0059] The first billing information may include at least one of the first order identification information, the installment benefits used for the processed order, the installment channel used for the processed order, and the number of installments used for the processed order.

[0060] In the application, the first order identification information includes but is not limited to user identification information, product identification information, merchant identification information and order number. The first billing information may also include order payment status, payment price and other information.

[0061] Step B2: Determine the target installment strategy information corresponding to the processed order based on the first order identification information.

[0062] Optionally, the fourth correspondence between the first order identification information and the target installment strategy information corresponding to the processed order can be stored in the installment strategy decision system, or can be sent by the cashier platform together with the first bill information.

[0063] Step B3: determining a first conversion result of the target installment strategy information corresponding to the processed order based on the first billing information and the target installment strategy information corresponding to the processed order.

[0064] In practice, step B3 may be performed by determining the first installment strategy information corresponding to the processed order based on the installment benefits used by the processed order, the installment channel used by the processed order, and the number of installments used by the processed order in the first billing information. Thus, by calculating the degree of match between the first installment strategy information and the target installment strategy information corresponding to the processed order, a first conversion result of the target installment strategy information corresponding to the processed order is determined.

[0065] For example, if the first installment strategy information completely matches the target installment strategy information corresponding to the processed order, that is, the two are completely consistent, it can be determined that the first conversion result of the target installment strategy information corresponding to the processed order is 100%.

[0066] Step B4: determining the decision accuracy of the installment strategy decision system on the installment method of the processed order based on the first conversion result and the preset effect evaluation strategy.

[0067] The preset effect evaluation strategy may include a preset accuracy requirement and a fifth corresponding relationship between the first conversion result and the decision accuracy.

[0068] Alternatively, the first conversion result can be used as the decision accuracy, or the decision accuracy corresponding to the first conversion result can be determined based on the fifth corresponding relationship described above. This allows determining whether the decision accuracy of the staging strategy decision system meets the preset accuracy requirements, and further determining whether the multi-objective prediction model needs to be revised.

[0069] In this embodiment, after the order processing is completed, the conversion status of the target installment strategy information corresponding to the processed order is determined based on the bill information fed back by the cash register platform, thereby determining the decision accuracy of the installment strategy decision system on the installment method of the processed order, providing a data basis for timely discovering the decision errors of the installment strategy decision system.

[0070] In one embodiment, when the decision accuracy of the installment strategy decision system on the installment method of the processed order does not meet the preset accuracy requirement, the multi-objective prediction model can be optimized according to the following steps C1 to C5.

[0071] Step C1: When the decision accuracy does not meet the preset accuracy requirement, obtain the second billing information of the target order.

[0072] The second billing information may include at least one of the second order identification information, the installment benefits used for the target order, the installment channel used for the target order, and the number of installments used for the target order. The target order is an order processed by the installment strategy decision system, and all installment strategy information has been sent to the cash register platform.

[0073] In the application, the second order identification information includes, but is not limited to, user identification information, product identification information, merchant identification information, and order number. The target order can be an order of a target user, who is a beta tester of the installment strategy decision system. For such users, the installment strategy decision system will output all installment strategy information. The user can select the most ideal installment strategy information, namely the target installment strategy information, for payment, or select other installment strategy information for payment. The user's order information and corresponding billing information can be used to optimize and improve the installment strategy decision system.

[0074] Step C2: Determine multiple installment strategy information corresponding to the target order based on the second order identification information.

[0075] Each staging strategy information is composed of at least one staging method in the prediction target.

[0076] For example, when the prediction target includes installment benefits, installment channels and installment periods, and the installment method includes sub-installment benefits, sub-installment channels and sub-installment periods, if the predicted value of the target order under interest-free is 10%, the predicted value under discount is 15%, the predicted value under red envelope is 25%, the predicted value under credit card is 15%, the predicted value under credit purchase is 10%, the predicted value under Huabei is 25%, the predicted value under 3 periods is 5%, the predicted value under 6 periods is 15%, and the predicted value under 12 periods is 30%, then, by freely combining the installment methods in each prediction target, 27 types of installment strategy information can be obtained, which will not be repeated here. Alternatively, by sorting the predicted values ​​corresponding to the installment methods in each prediction target in order from large to small, and combining them horizontally, three installment strategy information of red envelope, Huabei and 12 periods, discount, credit card and 6 periods, and interest-free, credit purchase and 3 periods are formed.

[0077] Step C3: Determine a second conversion result of the installment strategy information corresponding to the target order based on the second billing information and the multiple installment strategy information corresponding to the target order.

[0078] In practice, step C3 may be performed by determining the second installment strategy information corresponding to the target order based on the installment benefits used by the target order, the installment channel used by the target order, and the number of installments used by the target order in the second billing information. Thus, by calculating the degree of match between the second installment strategy information and the multiple installment strategy information corresponding to the target order, the second conversion result of the installment strategy information corresponding to the target order is determined.

[0079] For example, if the second installment strategy information completely matches a certain installment strategy information corresponding to the target order, that is, the two are completely consistent, then it can be determined that the second conversion result of the target order under the installment strategy information is 100%.

[0080] Step C4: when the second conversion result does not match the target installment strategy information corresponding to the target order, determining the decision error of the installment method for the target order.

[0081] Continuing with the example in step C2, the target installment strategy information corresponding to the target order is red envelopes, Huabei, and 12 installments. If, after calculation in step C3, it is determined that the second conversion result of the target order under red envelopes, Huabei, and 12 installments is 100%, then it can be determined that the second conversion result matches the target installment strategy information corresponding to the target order, thereby determining that there is no decision error in the multi-objective prediction model and no optimization processing is required. If, after calculation in step C3, it is determined that the second conversion result of the target order under discounts, Huabei, and 12 installments is 100%, then it can be determined that the second conversion result does not match the target installment strategy information corresponding to the target order, thereby determining that there is a decision error in the multi-objective prediction model, specifically, a decision error in the prediction of the installment benefit.

[0082] Step C5: According to the decision error, the model parameters of the multi-objective prediction model are corrected to obtain an optimized multi-objective prediction model.

[0083] In this step, if there is a decision error for any prediction target, the model parameters corresponding to the prediction target in the multi-target prediction model are corrected to make the model prediction result consistent with the actual more ideal staging strategy information.

[0084] In this embodiment, through real-time data processing and feedback mechanism, when the decision accuracy of the installment strategy decision system on the installment method of the processed order does not meet the preset accuracy requirements, the decision error of the installment strategy decision system can be determined based on the difference between the installment strategy information of the internal test user, the installment strategy information actually selected by the user and the target installment strategy information recommended by the installment strategy decision system. Based on the decision error, the model parameters of the multi-objective prediction model are corrected so that the model prediction results are consistent with the actual more ideal installment strategy information, so as to ensure the decision accuracy of the installment strategy decision system, realize the self-update effect of the algorithm strategy, ensure the effectiveness and flexibility of the strategy, and enhance the adaptability of the installment strategy decision system.

[0085] In one embodiment, determining target installment strategy information corresponding to the pending order based on the predicted value (i.e., S206) can be performed as follows: Steps E1 and E2: Step E1: Determine the calculated values ​​of the installment business target of the order to be processed under different installment methods based on the predicted value and the first order information.

[0086] The installment business objectives may include transaction volume and / or revenue. In the application, the installment business objectives of the installment strategy decision system can be set as transaction volume or revenue; or the installment business objectives of the installment strategy decision system can be set as transaction volume and revenue, and the corresponding order ratios are set for each. When a new order enters the installment strategy decision system, the installment business objectives of the current order are determined based on the preset order ratio.

[0087] Step E2: Determine the target installment strategy information corresponding to the pending order based on the calculated value and the preset installment strategy determination method.

[0088] In this embodiment, based on the predicted values ​​of the pending order under different forecast targets and different installment methods and the first order information, the calculated values ​​of the installment business targets for the pending order under different installment methods are determined. Based on the calculated values ​​and the preset installment strategy determination method, the target installment strategy information corresponding to the pending order is determined. This technical solution incorporates the influencing factor of the installment business target into the determination of the target installment strategy information, ensuring that the determined installment strategy information is correlated with the installment business target, facilitating the determination of installment strategy information that better meets the expectations of users of the installment strategy decision system (such as banks, third-party payment institutions, etc.).

[0089] In one embodiment, the installment business target may include transaction volume, and the first order information may include the first order amount. Thus, determining the calculated value of the installment business target for the pending order under different installment methods based on the predicted value and the first order information (i.e., step E1) can be performed by summing the predicted values ​​of the pending order under different installment methods for different predicted targets to obtain a summed result. Thus, the calculated value of the transaction volume for the pending order is determined based on the first order amount of the pending order and the summed result.

[0090] Optionally, when the multi-objective prediction model is a CVR model, the predicted value of the pending order under different prediction targets and different installment methods, the first order amount and the calculated value of the transaction volume The relationship between them is shown in formula (1).

[0091] , (1) Among them, channel (installment channel, recorded as ) and offer (installment benefits, remember ) as the scoring candidate set for the staging strategy decision system At the same time, drill down estimates exist (Installation period, such as 3 / 6 / 12 / 24 periods) CVR model score. The estimated abstraction is: the first order amount Weighted overall (CVR summary of all dimensions), completed Calculation of the target.

[0092] In one embodiment, the installment business objective may include revenue, and the first order information may include the first order amount, the first transaction type, and supported sub-installment channels. Thus, based on the predicted value and the first order information, the calculated values ​​of the installment business objective for the pending order under different installment methods (i.e., step E1) can be determined by determining, for each supported sub-installment channel, the fee rate and revenue coefficient corresponding to the first transaction type of the pending order. Consequently, the calculated revenue value for the pending order is determined based on the fee rate, revenue coefficient, and the calculated transaction volume of the pending order.

[0093] In the application, the first transaction type may include online shopping, insurance, merchant code shopping, personal code shopping, etc. The first transaction type may be a funding scenario .

[0094] Based on formula (1), the rate , profit coefficient , calculated value of the transaction volume of pending orders and the calculated value of the revenue of pending orders The relationship between them is shown in formula (2).

[0095] , (2) Among them, different installment channels (Remember to do ), investment scenarios (Remember to do ) and profit coefficient Different, so in On this basis, we further calculate the benefits under different circumstances.

[0096] In one embodiment, the installment business objectives may include transaction volume and revenue, and the first order information may include the first order amount, the first order creation time, and / or third order identification information. Thus, before determining the calculated values ​​of the installment business objectives for the pending order under different installment methods based on the predicted value and the first order information (i.e., step E1), the installment business objectives corresponding to the pending order can be determined based on the first order information and a preset installment business objective allocation strategy.

[0097] The third order identification information may include user identification information, product identification information, merchant identification information, order number, etc.

[0098] In this embodiment, after determining the installment business target corresponding to the pending order, the calculated value of the transaction volume of the pending order can be calculated with reference to the above formula (1), or the calculated value of the revenue of the pending order can be calculated with reference to the above formula (1) and formula (2).

[0099] In one embodiment, the preset installment business target allocation strategy includes at least one of the following: a first correspondence between the first order creation time and the installment business target; a second correspondence between the first order amount and the installment business target; and a third correspondence between the third order identification information and the installment business target.

[0100] Optionally, the first correspondence between the first order creation time and the installment business objective can be a correspondence between the creation time of each order within a specified time period and the installment business objective. For example, if the specified time period is 2 hours, the installment business objective for orders created within the first hour can be set to transaction volume, and the installment business objective for orders created within the second hour can be set to revenue.

[0101] The second correspondence between the first order amount and the installment business objective can be to set the installment business objective for orders exceeding a preset amount threshold as transaction volume (or revenue), and to set the installment business objective for orders less than or equal to the preset amount threshold as revenue (or transaction volume). Alternatively, the installment business objective for pending orders can be determined by calculating the order amounts of orders corresponding to the two installment business objectives of transaction volume and revenue, respectively, to balance the order amounts of each installment business objective.

[0102] The third correspondence between the third order identification information and the installment business goal may include a correspondence between the third order identification information and the transaction volume or revenue, so that the installment business goal corresponding to the pending order can be determined based on the third correspondence and the third order identification information.

[0103] Optionally, the preset installment business target allocation strategy may include the order ratio of each installment business target, so that after a new order enters the installment strategy decision system, the installment business target of the current order is determined based on the preset order ratio.

[0104] In one embodiment, based on the calculated value and a preset installment strategy determination method, the target installment strategy information corresponding to the pending order is determined (i.e., step E2), which can be performed as follows: steps F1 to F3: In step F1, according to a preset weighting strategy, the forecast values ​​of the pending orders under different forecast targets and different installment methods are weighted to obtain the weighted forecast values.

[0105] It is understood that weighting can, on the one hand, widen the gaps between predicted values, and on the other hand, can, to a certain extent, correct the decision-making errors of the multi-objective prediction model. In practice, the preset weighting strategies may include weighting strategies based on predicted values, weighting strategies based on installment channels, weighting strategies based on the number of installments, etc.

[0106] Alternatively, a weighting strategy based on predicted values ​​can be one where the predicted value is positively correlated with the weight, i.e., the larger the predicted value, the larger the weight assigned. A weighting strategy based on installment channels can be one where the channel weighting is based on the installment method corresponding to the maximum predicted value under the installment channel. A weighting strategy based on the number of installments can be one where the installment method corresponding to the maximum predicted value under the installment number is weighted by the installment method.

[0107] Step F2, based on the weighted forecast value, the calculated values ​​of the installment business objectives of the pending orders under different installment methods and the first multi-objective solution method, determines the target installment method of the pending orders in the forecast target.

[0108] Optionally, the first multi-objective solution method can be sorting, large-scale linear programming, or other methods. In practice, if the installment business objective is transaction volume, the weighted forecast values ​​can be directly sorted or solved using a large-scale linear programming method to determine the target installment strategy for pending orders under each forecast objective. In this case, transaction volume serves only as a reference; higher transaction volume indicates a higher likelihood of recommending a more optimal installment strategy for the pending order.

[0109] If the phased business goal is profit, then a large-scale linear programming solution can be performed on the weighted forecast values ​​and the calculated values ​​of the profit of the pending orders to obtain the target phased method for the pending orders under each forecast goal.

[0110] Step F3: Determine the target installment strategy information corresponding to the pending order based on the target installment method.

[0111] In this embodiment, the influencing factor of the installment business goal is introduced in the process of determining the target installment strategy information, so that the determined installment strategy information is related to the installment business goal, which is conducive to determining the installment strategy information that is more in line with the expectations of the users of the installment strategy decision system (such as banks, third-party payment institutions, etc.).

[0112] In one embodiment, based on the calculated value and the preset installment strategy determination method, the target installment strategy information corresponding to the pending order is determined (i.e., step E2), which can be performed as follows: based on the calculated value of the installment business target of the pending order under different installment methods and the second multi-objective solution method, the target installment method of the pending order in the prediction target is determined, and thus based on the target installment method, the target installment strategy information corresponding to the pending order is determined.

[0113] Optionally, the second multi-objective solution method can be a sorting method, a large-scale linear programming method, or the like. In implementation, the calculated values ​​of the phased business objectives of the pending orders under different phased methods can be directly sorted or solved using a large-scale linear programming method to obtain the target phased method for the pending orders under each forecast target.

[0114] In this embodiment, the influencing factor of the installment business goal is introduced in the process of determining the target installment strategy information, so that the determined installment strategy information is related to the installment business goal, which is conducive to determining the installment strategy information that is more in line with the expectations of the users of the installment strategy decision system (such as banks, third-party payment institutions, etc.).

[0115] Figure 4 is a schematic flow chart of a payment processing method according to another embodiment of this specification. In this embodiment, the payment processing method is applied to Figure 1 The cash register platform 120 shown in FIG. Figure 4 As shown, the method may include: S402, obtaining first order information of a pending order, and receiving target installment strategy information corresponding to the pending order sent by the installment strategy decision system.

[0116] The target installment strategy information is obtained by inputting the first order information of a pending order into a pre-trained multi-objective prediction model, outputting the predicted values ​​of the pending order under different prediction objectives and different installment methods, and then determining the target installment strategy information corresponding to the pending order based on the predicted values. The target installment strategy information is composed of the target installment methods in the prediction objectives.

[0117] Optionally, the first order information may include at least one of the order amount, order creation time, order identification information (such as user identification information, product identification information, merchant identification information, order number, etc.), transaction type, supported sub-installment benefits, supported sub-installment channels, and supported sub-installment period numbers.

[0118] The first order information of a pending order can be sent by the downstream traffic pool to the cash register platform. It is understood that the orders in the downstream traffic pool include not only pending orders that support installment payments, but also orders that do not support installment payments. For orders that do not support installment payments, the downstream traffic pool will directly send them to the cash register platform for processing.

[0119] S404: Determine multiple candidate payment methods corresponding to the order to be processed based on the first order information.

[0120] During implementation, the cash register platform will determine multiple candidate payment methods corresponding to the pending order based on the first order information obtained and its own decision-making strategy. Optionally, the multiple candidate payment methods include but are not limited to balance payment, bank card payment, and installment payment.

[0121] S406: Determine recommended payment method information based on the multiple candidate payment methods and target installment strategy information.

[0122] S408: Display recommended payment method information.

[0123] In the application, users can select a payment method to pay for pending orders based on the payment method information displayed by the cash register platform.

[0124] By adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the cash register platform, and the target installment strategy information corresponding to the pending order sent by the installment strategy decision system is received. The target installment strategy information is obtained in the following way: the first order information of the pending order is input into a pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output, so as to determine the target installment strategy information corresponding to the pending order based on the predicted value. Thus, the cash register platform can determine the multiple candidate payment methods corresponding to the pending order based on the first order information, determine the recommended payment method information based on the multiple candidate payment methods and the target installment strategy information, and display the recommended payment method information. It can be seen that in this technical solution, since there is no need for the cash register platform to determine the installment strategy, the decision complexity of the cash register platform is reduced. Moreover, since the received target installment strategy information is a prediction result of the user's payment for the pending order using various installment methods of various prediction targets based on the first order information of the pending order using a multi-objective prediction model, targeted, accurate and efficient installment strategy decisions for the pending order are achieved, which is conducive to the cash register platform giving priority to exposing the installment payment method and the corresponding installment strategy information in limited exposure positions, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments.

[0125] In one embodiment, determining recommended payment method information based on multiple candidate payment methods and target installment strategy information (i.e., S406) can be performed by weighting the payment methods corresponding to the target installment strategy information among the multiple candidate payment methods based on the target installment strategy information to obtain multiple weighted candidate payment methods. Thus, the recommended payment method information is determined based on the top N candidate payment methods among the weighted multiple candidate payment methods.

[0126] Wherein, N is an integer greater than or equal to 1.

[0127] This embodiment uses big data analysis and machine learning algorithms to accurately identify users and transaction behaviors with high-value potential from massive transaction data. That is, it uses the installment strategy decision system to provide targeted installment strategy decisions for orders that support installment payments, and transmits the target installment strategy information to the upstream cash register platform, so that the cash register platform can perform weighted processing on the payment methods corresponding to the target installment strategy information among multiple candidate payment methods based on the target installment strategy information, thereby ensuring that this payment method is given priority when displayed, thereby improving its conversion rate and transaction success rate.

[0128] Figure 5 This is a schematic swimlane diagram of a payment processing method according to an embodiment of this specification. Through the interaction between the downstream traffic pool, the installment strategy decision system and the cash register platform, targeted, accurate and efficient installment strategy decisions can be made for pending orders, which is beneficial for the cash register platform to prioritize the installment payment method and the corresponding installment strategy information in limited exposure locations, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments. Figure 5 As shown, the following steps S5.1-S5.8 are included: S5.1, the downstream traffic pool sends the first order information of the pending order to the installment strategy decision system and the cash register platform respectively.

[0129] Among them, pending orders are orders that support installment payment.

[0130] S5.2, the installment strategy decision system inputs the first order information into a pre-trained multi-objective prediction model, and outputs the predicted values ​​of the pending order under different installment methods with different prediction objectives.

[0131] The predicted value represents the predicted result of payment for pending orders using the installment method for the predicted target. The predicted target may include installment benefits, installment channels, and the number of installments. The installment method may include sub-installment benefits, sub-installment channels, and the number of sub-installments.

[0132] S5.3, the installment strategy decision system determines the calculated values ​​of the installment business target of the pending order under different installment methods based on the predicted value and the first order information.

[0133] Wherein, the phased business objectives include transaction volume and / or revenue. Based on the different phased business objectives, the transaction volume of pending orders can be calculated by referring to the above formula (1), or the revenue of pending orders can be calculated by referring to the above formulas (1) and (2).

[0134] S5.4, the installment strategy decision system determines the target installment strategy information corresponding to the pending order based on the calculated value and the preset installment strategy determination method.

[0135] The specific implementation process of this step can refer to the relevant embodiment of the above-mentioned step E2, which will not be described in detail here. The target staging strategy information is composed of the target staging method in the predicted target.

[0136] S5.5, the installment strategy decision system sends the target installment strategy information to the cash register platform.

[0137] S5.6. The cash register platform determines multiple candidate payment methods corresponding to the pending order based on the first order information.

[0138] It should be noted that the execution order of S5.6 and S5.2-S5.5 is not limited in this embodiment. For example, in addition to the execution order of S5.2-S5.5 first and then S5.6 as listed in this embodiment, S5.6 can also be executed first and then S5.2-S5.5, or S5.6 and S5.2-S5.5 can be executed simultaneously.

[0139] S5.7, the cash register platform determines recommended payment method information based on multiple candidate payment methods and target installment strategy information.

[0140] S5.8, the cash register platform displays the recommended payment method information.

[0141] The specific process of S5.1-S5.8 has been described in detail in the above embodiment and will not be repeated here.

[0142] Using the technical solutions of one or more embodiments of this specification, an installment strategy decision system obtains first order information for a pending order, inputs the first order information into a pre-trained multi-objective prediction model, and outputs predicted values ​​for the pending order under different installment methods for different prediction targets. Based on the predicted values, the target installment strategy information corresponding to the pending order is determined, and the target installment strategy information is sent to a cash register platform. The cash register platform determines and displays recommended payment method information based on the received first order information and target installment strategy information. This technical solution can utilize the multi-objective prediction model to specifically predict the user's payment results for the pending order using various installment methods for each prediction target based on the first order information of the pending order, and then output the target installment strategy information determined based on the prediction results to the cash register platform. This enables targeted, accurate, and efficient installment strategy decisions for pending orders, facilitating the cash register platform to prioritize the installment payment method and corresponding installment strategy information in limited exposure locations, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments. Moreover, since the cash register platform does not need to determine the installment strategy, the decision-making complexity of the cash register platform is reduced.

[0143] In summary, specific embodiments of the present subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing may be advantageous.

[0144] The above is a payment processing method provided in one or more embodiments of this specification. Based on the same idea, one or more embodiments of this specification also provide a payment processing device.

[0145] Figure 6 This is a schematic block diagram of a payment processing device according to an embodiment of this specification, which is applied to the installment strategy decision system. Please refer to Figure 6 , the payment processing device may include: A first obtaining module 610 is used to obtain first order information of an order to be processed; Model prediction module 620 is configured to input the first order information into a pre-trained multi-objective prediction model and output predicted values ​​for the pending order under different prediction objectives and different installment methods; the predicted values ​​are used to represent the predicted results of payment for the pending order using the installment method of the prediction objective; the prediction objective includes at least one of installment benefits, installment channels, and the number of installments; and the installment method includes at least one of sub-installment benefits, sub-installment channels, and the number of sub-installments; The determination and sending module 630 is used to determine the target installment strategy information corresponding to the pending order based on the predicted value, and send the target installment strategy information to the cash register platform; the target installment strategy information is composed of the target installment method in the predicted target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

[0146] In one embodiment, the payment processing device further includes: The second acquisition module is configured to acquire sample order information and sample billing information for a plurality of sample orders; the sample order information includes at least one of the following: the sample order amount, the sample order creation time, the sample order identification information, the sample transaction type, the supported sub-installment benefits, the supported sub-installment channels, and the supported number of sub-installment periods; and the sample billing information includes the installment benefits used by the sample order, the installment channels used by the sample order, and the number of installment periods used by the sample order; The model training module is used to input sample order information and sample bill information into the multi-objective prediction model to be trained for iterative training to obtain the trained multi-objective prediction model.

[0147] In one embodiment, the model training module includes: A prediction unit is used to predict, in the multi-objective prediction model to be trained, based on the sample order information, a sample prediction value of payment for the sample order using each installment method for each prediction objective; A first determining unit is configured to determine, for each sample order, a loss value for predicting the sample order using the multi-objective prediction model based on the sample prediction value of the sample order in each installment mode, the sample bill information, and a loss function corresponding to the multi-objective prediction model; The model parameter iteration unit is used to iteratively adjust the model parameters according to the loss value until the iteration termination condition is met, stop training, and obtain the trained multi-objective prediction model.

[0148] In one embodiment, the payment processing device further includes: a receiving module, configured to receive first billing information of a processed order fed back by the cash register platform; the first billing information including at least one of first order identification information, installment benefits used for the processed order, installment channels used for the processed order, and number of installments used for the processed order; A first determining module is configured to determine target installment strategy information corresponding to the processed order based on the first order identification information; a second determining module, configured to determine a first conversion result of the target installment strategy information corresponding to the processed order based on the first billing information and the target installment strategy information corresponding to the processed order; The third determination module is used to determine the decision accuracy of the installment strategy decision system on the installment method of the processed order based on the first conversion result and a preset effect evaluation strategy.

[0149] In one embodiment, the payment processing device further includes: A third acquisition module is configured to acquire second billing information for the target order if the decision accuracy does not meet a preset accuracy requirement; the second billing information includes at least one of second order identification information, the installment benefit used for the target order, the installment channel used for the target order, and the number of installments used for the target order; the target order is an order processed by the installment strategy decision system, and all installment strategy information has been sent to the cash register platform; A fourth determining module is configured to determine, based on the second order identification information, a plurality of installment strategy information corresponding to the target order; each installment strategy information is composed of an installment method in at least one predicted target; A fifth determining module, configured to determine a second conversion result of the installment strategy information corresponding to the target order based on the second billing information and the multiple installment strategy information corresponding to the target order; a sixth determining module, configured to determine a decision error of the installment method for the target order when the second conversion result does not match the target installment strategy information corresponding to the target order; The model parameter correction module is used to correct the model parameters of the multi-objective prediction model according to the decision error to obtain an optimized multi-objective prediction model.

[0150] In one embodiment, the determining and sending module 630 includes: A second determining unit is configured to determine, based on the predicted value and the first order information, calculated values ​​of the installment business objectives of the pending order under different installment methods; the installment business objectives include transaction volume and / or revenue; The third determining unit is configured to determine target installment strategy information corresponding to the pending order based on the calculated value and a preset installment strategy determination method.

[0151] In one embodiment, the installment business target includes transaction volume; the first order information includes the first order amount; The second determining unit is specifically configured to: The forecast values ​​of the pending orders under different forecast targets and different installment methods are summed to obtain the sum result; The calculated value of the transaction volume of the pending order is determined based on the first order amount of the pending order and the sum result.

[0152] In one embodiment, the installment business goal includes revenue; the first order information further includes the first transaction type and supported sub-installment channels; The second determining unit is specifically configured to: For each supported sub-installment channel, determine the fee rate and profit coefficient corresponding to the first transaction type of the pending order; The calculated value of the profit of the pending order is determined based on the calculated value of the fee rate, the profit coefficient, and the trading volume of the pending order.

[0153] In one embodiment, the installment business objectives include transaction volume and revenue; the first order information also includes the first order creation time and / or third order identification information; The determination and sending module 630 further includes: The fourth determination unit is used to determine the installment business target corresponding to the pending order based on the first order information and the preset installment business target allocation strategy before determining the calculated value of the installment business target of the pending order under different installment methods based on the predicted value and the first order information.

[0154] In one embodiment, the preset phased business target allocation strategy includes at least one of the following: The first correspondence between the first order creation time and the installment business target; A second correspondence between the first order amount and the installment business target; The third correspondence between the third order identification information and the installment business target.

[0155] In one embodiment, the third determining unit is specifically configured to: According to the preset weighting strategy, the predicted value is weighted to obtain the weighted predicted value; Determine the target installment method for the pending order in the forecast target based on the weighted forecast value, the calculated values ​​of the installment business target of the pending order under different installment methods, and the first multi-objective solution method; Based on the target installment method, determine the target installment strategy information corresponding to the pending orders.

[0156] In one embodiment, the third determining unit is specifically configured to: Determine the target installment method for the pending orders in the forecast target based on the calculated values ​​of the installment business objectives of the pending orders under different installment methods and the second multi-objective solution method; Based on the target installment method, determine the target installment strategy information corresponding to the pending orders.

[0157] Adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the installment strategy decision system, the first order information is input into the pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output. The prediction target includes at least one of the installment rights, installment channels, and installment period numbers, and the installment method includes at least one of the sub-installment rights, sub-installment channels, and sub-installment period numbers. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. Thus, the target installment strategy information corresponding to the pending order is determined according to the predicted value, and the target installment strategy information is sent to the cash register platform. The target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information. As can be seen, this technical solution can utilize a multi-objective prediction model to specifically predict the user's payment results for the pending order using various installment methods for each prediction target based on the first order information of the pending order, thereby outputting the target installment strategy information determined based on the prediction results to the cash register platform, achieving targeted, accurate, and efficient installment strategy decisions for the pending order. This helps the cash register platform prioritize the installment payment method and corresponding installment strategy information in limited exposure locations, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments. In addition, since the cash register platform does not need to determine the installment strategy, the decision-making complexity of the cash register platform is reduced.

[0158] Figure 7 This is a schematic block diagram of a payment processing device according to another embodiment of this specification, which is applied to a cash register platform. Please refer to Figure 7 , the payment processing device may include: The acquisition and receiving module 710 is configured to acquire first order information of a pending order and receive target installment strategy information corresponding to the pending order from the installment strategy decision system. The target installment strategy information is obtained by inputting the first order information of the pending order into a pre-trained multi-objective prediction model, outputting predicted values ​​of the pending order under different prediction objectives and different installment methods, and determining the target installment strategy information corresponding to the pending order based on the predicted values. A seventh determining module 720 is configured to determine, based on the first order information, multiple candidate payment methods corresponding to the pending order; An eighth determination module 730 is configured to determine recommended payment method information based on multiple candidate payment methods and target installment strategy information; The display module 740 is used to display recommended payment method information.

[0159] In one embodiment, the eighth determination module 730 includes: a weighted processing unit configured to perform weighted processing on the payment methods corresponding to the target installment strategy information among the multiple candidate payment methods according to the target installment strategy information, thereby obtaining multiple candidate payment methods after weighted processing; The fifth determining unit is configured to determine recommended payment method information based on top N candidate payment methods among the plurality of candidate payment methods after weighted processing; N is an integer greater than or equal to 1.

[0160] By adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the cash register platform, and the target installment strategy information corresponding to the pending order sent by the installment strategy decision system is received. The target installment strategy information is obtained in the following way: the first order information of the pending order is input into a pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output, so as to determine the target installment strategy information corresponding to the pending order based on the predicted value. Thus, the cash register platform can determine the multiple candidate payment methods corresponding to the pending order based on the first order information, determine the recommended payment method information based on the multiple candidate payment methods and the target installment strategy information, and display the recommended payment method information. It can be seen that in this technical solution, since there is no need for the cash register platform to determine the installment strategy, the decision complexity of the cash register platform is reduced. Moreover, since the received target installment strategy information is a prediction result of the user's payment for the pending order using various installment methods of various prediction targets based on the first order information of the pending order using a multi-objective prediction model, targeted, accurate and efficient installment strategy decisions for the pending order are achieved, which is conducive to the cash register platform giving priority to exposing the installment payment method and the corresponding installment strategy information in limited exposure positions, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments.

[0161] Those skilled in the art should understand that the above-mentioned payment processing device can be used to implement the payment processing method described above, and the detailed description should be similar to the description of the method part above. To avoid redundancy, it will not be repeated here.

[0162] Based on the same idea, one or more embodiments of this specification also provide a payment processing device, such as Figure 8As shown. The payment processing device may vary greatly due to different configurations or performance, and may include one or more processors 801 and memory 802. The memory 802 may store one or more application programs or data. The memory 802 may be a temporary storage or a persistent storage. The application stored in the memory 802 may include one or more modules (not shown in the figure), each module may include a series of computer-executable instructions for the payment processing device. Furthermore, the processor 801 may be configured to communicate with the memory 802 to execute a series of computer-executable instructions in the memory 802 on the payment processing device. The payment processing device may also include one or more power supplies 803, one or more wired or wireless network interfaces 804, one or more input and output interfaces 805, and one or more keyboards 806.

[0163] In one embodiment, a payment processing device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the payment processing device, and the one or more programs are configured to be executed by one or more processors, including computer-executable instructions for performing the following: Get the first order information of pending orders; Inputting the first order information into a pre-trained multi-objective prediction model, and outputting predicted values ​​of the pending order under different prediction targets and different installment methods; the predicted values ​​are used to represent the predicted results of payment for the pending order using the installment method of the prediction target; the prediction target includes at least one of installment benefits, installment channels, and the number of installments; and the installment method includes at least one of sub-installment benefits, sub-installment channels, and the number of sub-installments; The target installment strategy information corresponding to the pending order is determined based on the predicted value, and the target installment strategy information is sent to the cash register platform; the target installment strategy information is composed of the target installment method in the predicted target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

[0164] Adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the installment strategy decision system, the first order information is input into the pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output. The prediction target includes at least one of the installment rights, installment channels, and installment period numbers, and the installment method includes at least one of the sub-installment rights, sub-installment channels, and sub-installment period numbers. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. Thus, the target installment strategy information corresponding to the pending order is determined according to the predicted value, and the target installment strategy information is sent to the cash register platform. The target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information. As can be seen, this technical solution can utilize a multi-objective prediction model to specifically predict the user's payment results for the pending order using various installment methods for each prediction target based on the first order information of the pending order, thereby outputting the target installment strategy information determined based on the prediction results to the cash register platform, achieving targeted, accurate, and efficient installment strategy decisions for the pending order. This helps the cash register platform prioritize the installment payment method and corresponding installment strategy information in limited exposure locations, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments. In addition, since the cash register platform does not need to determine the installment strategy, the decision-making complexity of the cash register platform is reduced.

[0165] In one embodiment, a payment processing device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the payment processing device, and the one or more programs are configured to be executed by one or more processors, including computer-executable instructions for performing the following: Obtaining first order information of a pending order, and receiving target installment strategy information corresponding to the pending order from an installment strategy decision system; the target installment strategy information is obtained by: inputting the first order information of the pending order into a pre-trained multi-objective prediction model, outputting predicted values ​​of the pending order under different prediction objectives and different installment methods; and determining the target installment strategy information corresponding to the pending order based on the predicted values; Determining, based on the first order information, multiple candidate payment methods corresponding to the pending order; Determine recommended payment method information based on multiple candidate payment methods and target installment strategy information; Displays recommended payment method information.

[0166] By adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the cash register platform, and the target installment strategy information corresponding to the pending order sent by the installment strategy decision system is received. The target installment strategy information is obtained in the following way: the first order information of the pending order is input into a pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output, so as to determine the target installment strategy information corresponding to the pending order based on the predicted value. Thus, the cash register platform can determine the multiple candidate payment methods corresponding to the pending order based on the first order information, determine the recommended payment method information based on the multiple candidate payment methods and the target installment strategy information, and display the recommended payment method information. It can be seen that in this technical solution, since there is no need for the cash register platform to determine the installment strategy, the decision complexity of the cash register platform is reduced. Moreover, since the received target installment strategy information is a prediction result of the user's payment for the pending order using various installment methods of various prediction targets based on the first order information of the pending order using a multi-objective prediction model, targeted, accurate and efficient installment strategy decisions for the pending order are achieved, which is conducive to the cash register platform giving priority to exposing the installment payment method and the corresponding installment strategy information in limited exposure positions, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments.

[0167] One or more embodiments of this specification further provide a storage medium storing one or more computer programs. The one or more computer programs include instructions that, when executed by an electronic device including multiple application programs, enable the electronic device to perform various processes of the payment processing method embodiment described above, and are specifically configured to perform: Get the first order information of pending orders; Inputting the first order information into a pre-trained multi-objective prediction model, and outputting predicted values ​​of the pending order under different prediction targets and different installment methods; the predicted values ​​are used to represent the predicted results of payment for the pending order using the installment method of the prediction target; the prediction target includes at least one of installment benefits, installment channels, and the number of installments; and the installment method includes at least one of sub-installment benefits, sub-installment channels, and the number of sub-installments; The target installment strategy information corresponding to the pending order is determined based on the predicted value, and the target installment strategy information is sent to the cash register platform; the target installment strategy information is composed of the target installment method in the predicted target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

[0168] Adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the installment strategy decision system, the first order information is input into the pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output. The prediction target includes at least one of the installment rights, installment channels, and installment period numbers, and the installment method includes at least one of the sub-installment rights, sub-installment channels, and sub-installment period numbers. The predicted value is used to characterize the predicted result of paying the pending order using the installment method of the prediction target. Thus, the target installment strategy information corresponding to the pending order is determined according to the predicted value, and the target installment strategy information is sent to the cash register platform. The target installment strategy information is composed of the target installment method in the prediction target, and is used by the cash register platform to determine the recommended payment method information based on the first order information and the target installment strategy information. As can be seen, this technical solution can utilize a multi-objective prediction model to specifically predict the user's payment results for the pending order using various installment methods for each prediction target based on the first order information of the pending order, thereby outputting the target installment strategy information determined based on the prediction results to the cash register platform, achieving targeted, accurate, and efficient installment strategy decisions for the pending order. This helps the cash register platform prioritize the installment payment method and corresponding installment strategy information in limited exposure locations, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments. In addition, since the cash register platform does not need to determine the installment strategy, the decision-making complexity of the cash register platform is reduced.

[0169] One or more embodiments of this specification further provide a storage medium storing one or more computer programs. The one or more computer programs include instructions that, when executed by an electronic device including multiple application programs, enable the electronic device to perform various processes of the payment processing method embodiment described above, and are specifically configured to perform: Obtaining first order information of a pending order, and receiving target installment strategy information corresponding to the pending order from an installment strategy decision system; the target installment strategy information is obtained by: inputting the first order information of the pending order into a pre-trained multi-objective prediction model, outputting predicted values ​​of the pending order under different prediction objectives and different installment methods; and determining the target installment strategy information corresponding to the pending order based on the predicted values; Determining, based on the first order information, multiple candidate payment methods corresponding to the pending order; Determine recommended payment method information based on multiple candidate payment methods and target installment strategy information; Displays recommended payment method information.

[0170] By adopting the technical solution of one or more embodiments of this specification, the first order information of the pending order is obtained through the cash register platform, and the target installment strategy information corresponding to the pending order sent by the installment strategy decision system is received. The target installment strategy information is obtained in the following way: the first order information of the pending order is input into a pre-trained multi-objective prediction model, and the predicted value of the pending order under different installment methods with different prediction targets is output, so as to determine the target installment strategy information corresponding to the pending order based on the predicted value. Thus, the cash register platform can determine the multiple candidate payment methods corresponding to the pending order based on the first order information, determine the recommended payment method information based on the multiple candidate payment methods and the target installment strategy information, and display the recommended payment method information. It can be seen that in this technical solution, since there is no need for the cash register platform to determine the installment strategy, the decision complexity of the cash register platform is reduced. Moreover, since the received target installment strategy information is a prediction result of the user's payment for the pending order using various installment methods of various prediction targets based on the first order information of the pending order using a multi-objective prediction model, targeted, accurate and efficient installment strategy decisions for the pending order are achieved, which is conducive to the cash register platform giving priority to exposing the installment payment method and the corresponding installment strategy information in limited exposure positions, thereby encouraging users to make installment payments for pending orders, thereby increasing the transaction volume and revenue of installment payments.

[0171] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, 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.

[0172] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing one or more embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0173] Those skilled in the art will appreciate that one or more embodiments of this specification may be provided as a method, system, or computer program product. Thus, one or more embodiments of this specification may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, one or more embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0174] One or more embodiments of this specification are described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0175] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0176] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0177] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0178] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0179] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology for information storage. 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 RAM (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 cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0180] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0181] One or more embodiments of this specification may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communications network. In distributed computing environments, program modules may be located in local and remote computer storage media, including storage devices.

[0182] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0183] The foregoing is merely one or more embodiments of this specification and is not intended to limit this application. It will be apparent to those skilled in the art that various modifications and variations may be made to one or more embodiments of this specification. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of this specification shall be included within the scope of the claims of one or more embodiments of this specification.

Claims

1. A payment processing method, applied to an installment strategy decision system, comprising: Get the first order information of pending orders; Inputting the first order information into a pre-trained multi-objective prediction model, and outputting predicted values ​​of the pending order under different prediction targets and different installment methods; the predicted values ​​are used to represent the predicted results of payment for the pending order using the installment method of the prediction target; the prediction target includes at least one of installment benefits, installment channels, and installment period number; the installment method includes at least one of sub-installment benefits, sub-installment channels, and sub-installment period number; Determine the target installment strategy information corresponding to the pending order according to the predicted value, and send the target installment strategy information to the cash register platform; The target installment strategy information is composed of the target installment method in the predicted target, and is used by the cashier platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

2. The method according to claim 1, further comprising: Obtaining sample order information and sample billing information for multiple sample orders; the sample order information includes at least one of the sample order amount, sample order creation time, sample order identification information, sample transaction type, supported sub-installment benefits, supported sub-installment channels, and supported sub-installment period numbers; the sample billing information includes the installment benefits used in the sample orders, the installment channels used in the sample orders, and the installment period numbers used in the sample orders; The sample order information and the sample bill information are input into the multi-objective prediction model to be trained for iterative training to obtain a trained multi-objective prediction model.

3. The method according to claim 2, wherein the step of inputting the sample order information and the sample bill information into a multi-objective prediction model to be trained for iterative training to obtain a trained multi-objective prediction model comprises: In the multi-objective prediction model to be trained, based on the sample order information, a sample prediction value of paying the sample order using each of the installment methods of each of the prediction objectives is predicted; For each of the sample orders, determining a loss value for predicting the sample order using the multi-objective prediction model based on the sample prediction value of the sample order under each of the installment methods, the sample bill information, and a loss function corresponding to the multi-objective prediction model; According to the loss value, the model parameters are iteratively adjusted until the iteration termination condition is met, and the training is stopped to obtain the trained multi-objective prediction model.

4. The method according to claim 1, further comprising: Receiving first billing information of the processed order fed back by the cash register platform; The first billing information includes at least one of the first order identification information, the installment benefits used for the processed order, the installment channel used for the processed order, and the number of installments used for the processed order; Determining target installment strategy information corresponding to the processed order based on the first order identification information; determining a first conversion result of the target installment strategy information corresponding to the processed order based on the first billing information and the target installment strategy information corresponding to the processed order; The decision accuracy of the installment method of the processed order made by the installment strategy decision system is determined based on the first conversion result and a preset effect evaluation strategy.

5. The method according to claim 4, further comprising: If the decision accuracy does not meet the preset accuracy requirement, obtaining second billing information of the target order; The second billing information includes at least one of second order identification information, installment benefits used for the target order, installment channels used for the target order, and the number of installments used for the target order; the target order is an order processed by the installment strategy decision system, and all installment strategy information is sent to the cash register platform; Determining, based on the second order identification information, a plurality of installment strategy information corresponding to the target order; each of the installment strategy information is composed of the installment method in at least one of the predicted targets; determining a second conversion result of the installment strategy information corresponding to the target order based on the second billing information and the multiple installment strategy information corresponding to the target order; If the second conversion result does not match the target installment strategy information corresponding to the target order, determining a decision error for the installment method of the target order; According to the decision error, the model parameters of the multi-objective prediction model are corrected to obtain an optimized multi-objective prediction model.

6. The method according to claim 1, wherein determining target installment strategy information corresponding to the pending order based on the predicted value comprises: Determining calculated values ​​of the installment business target of the pending order under the different installment methods based on the predicted value and the first order information; The phased business objectives include transaction volume and / or revenue; Based on the calculated value and a preset installment strategy determination method, target installment strategy information corresponding to the pending order is determined.

7. The method according to claim 6, wherein the installment business target includes transaction volume; the first order information includes the first order amount; Determining, based on the predicted value and the first order information, calculated values ​​of the installment business target of the pending order under the different installment methods includes: Summing the forecast values ​​of the pending orders under different forecast targets and different installment methods to obtain a summation result; A calculated value of the transaction volume of the pending order is determined according to the first order amount of the pending order and the summation result.

8. The method according to claim 7, wherein the installment business goal includes revenue; the first order information further includes a first transaction type and supported sub-installment channels; Determining, based on the predicted value and the first order information, calculated values ​​of the installment business target of the pending order under the different installment methods includes: Determining, for each of the supported sub-installment channels, a fee rate and a profit coefficient corresponding to the first transaction type of the pending order; The calculated value of the profit of the pending order is determined according to the fee rate, the profit coefficient and the calculated value of the transaction volume of the pending order.

9. The method according to claim 8, wherein the installment business target includes the transaction volume and the revenue; the first order information further includes the first order creation time and / or third order identification information; Before determining the calculated values ​​of the installment business target of the pending order under the different installment methods based on the predicted value and the first order information, the method further includes: Based on the first order information and a preset installment business target allocation strategy, the installment business target corresponding to the pending order is determined.

10. The method according to claim 9, wherein the preset phased business target allocation strategy includes at least one of the following: a first correspondence between the first order creation time and the installment business goal; a second corresponding relationship between the first order amount and the installment business target; A third corresponding relationship between the third order identification information and the installment business target.

11. The method according to claim 6, wherein determining target installment strategy information corresponding to the pending order based on the calculated value and a preset installment strategy determination method comprises: Performing weighted processing on the predicted value according to a preset weighting strategy to obtain a weighted predicted value; Determining a target installment method for the pending order in the forecast target based on the weighted forecast value, the calculated values ​​of the installment business target of the pending order under the different installment methods, and the first multi-objective solution method; Based on the target installment method, target installment strategy information corresponding to the pending order is determined.

12. The method according to claim 6, wherein determining target installment strategy information corresponding to the pending order based on the calculated value and a preset installment strategy determination method comprises: Determining a target installment method for the pending order in the forecast target based on the calculated values ​​of the installment business objectives of the pending order under the different installment methods and the second multi-objective solution method; Based on the target installment method, target installment strategy information corresponding to the pending order is determined.

13. A payment processing method, applied to a cash register platform, comprising: Acquire first order information of a pending order, and receive target installment strategy information corresponding to the pending order sent by an installment strategy decision system; The target installment strategy information is obtained by: inputting the first order information of the pending order into a pre-trained multi-objective prediction model, outputting the predicted values ​​of the pending order under different prediction targets and different installment methods; and determining the target installment strategy information corresponding to the pending order based on the predicted values; Determining, based on the first order information, multiple candidate payment methods corresponding to the pending order; Determining recommended payment method information based on the multiple candidate payment methods and the target installment strategy information; The recommended payment method information is displayed.

14. The method according to claim 13, wherein determining the recommended payment method information based on the multiple candidate payment methods and the target installment strategy information comprises: performing weighted processing on the payment methods corresponding to the target installment strategy information among the multiple candidate payment methods according to the target installment strategy information to obtain multiple weighted candidate payment methods; Recommended payment method information is determined based on top N candidate payment methods among the multiple candidate payment methods after the weighted processing; N is an integer greater than or equal to 1.

15. A payment processing device, applied to an installment strategy decision system, comprising: A first acquisition module, used to acquire first order information of a pending order; a model prediction module, configured to input the first order information into a pre-trained multi-objective prediction model and output predicted values ​​of the pending order under different prediction objectives and different installment methods; the predicted values ​​are used to represent the predicted results of payment for the pending order using the installment method of the prediction objective; the prediction objective includes at least one of installment benefits, installment channels, and the number of installments; and the installment method includes at least one of sub-installment benefits, sub-installment channels, and the number of sub-installments; a determination and sending module, configured to determine target installment strategy information corresponding to the pending order based on the predicted value, and send the target installment strategy information to a cash register platform; The target installment strategy information is composed of the target installment method in the predicted target, and is used by the cashier platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

16. A payment processing device comprising: processor; as well as a memory arranged to store computer-executable instructions which, when executed, cause the processor to: Get the first order information of pending orders; Inputting the first order information into a pre-trained multi-objective prediction model, and outputting predicted values ​​of the pending order under different prediction targets and different installment methods; the predicted values ​​are used to represent the predicted results of payment for the pending order using the installment method of the prediction target; the prediction target includes at least one of installment benefits, installment channels, and installment period number; the installment method includes at least one of sub-installment benefits, sub-installment channels, and sub-installment period number; Determine the target installment strategy information corresponding to the pending order according to the predicted value, and send the target installment strategy information to the cash register platform; The target installment strategy information is composed of the target installment method in the predicted target, and is used by the cashier platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

17. A storage medium for storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the following process: Get the first order information of pending orders; Inputting the first order information into a pre-trained multi-objective prediction model, and outputting predicted values ​​of the pending order under different prediction targets and different installment methods; the predicted values ​​are used to represent the predicted results of payment for the pending order using the installment method of the prediction target; the prediction target includes at least one of installment benefits, installment channels, and installment period number; the installment method includes at least one of sub-installment benefits, sub-installment channels, and sub-installment period number; The target installment strategy information corresponding to the pending order is determined based on the predicted value, and the target installment strategy information is sent to the cashier platform; the target installment strategy information is composed of the target installment method in the predicted target, and is used by the cashier platform to determine the recommended payment method information based on the first order information and the target installment strategy information.

Images