CN113627900A - Model training method, device and storage medium - Google Patents
Model training method, device and storage medium Download PDFInfo
- Publication number
- CN113627900A CN113627900A CN202110913896.2A CN202110913896A CN113627900A CN 113627900 A CN113627900 A CN 113627900A CN 202110913896 A CN202110913896 A CN 202110913896A CN 113627900 A CN113627900 A CN 113627900A
- Authority
- CN
- China
- Prior art keywords
- target
- model
- trained
- routing
- attribute data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 61
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000004891 communication Methods 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 11
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 238000005259 measurement Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 8
- 238000012423 maintenance Methods 0.000 claims description 7
- 238000013473 artificial intelligence Methods 0.000 abstract description 7
- 238000005516 engineering process Methods 0.000 description 9
- 238000013461 design Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/02—Payment architectures, schemes or protocols involving a neutral party, e.g. certification authority, notary or trusted third party [TTP]
- G06Q20/027—Payment architectures, schemes or protocols involving a neutral party, e.g. certification authority, notary or trusted third party [TTP] involving a payment switch or gateway
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/08—Payment architectures
- G06Q20/085—Payment architectures involving remote charge determination or related payment systems
- G06Q20/0855—Payment architectures involving remote charge determination or related payment systems involving a third party
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/08—Payment architectures
- G06Q20/10—Payment architectures specially adapted for electronic funds transfer [EFT] systems; specially adapted for home banking systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q20/403—Solvency checks
- G06Q20/4037—Remote solvency checks
Abstract
The application relates to the technical field of artificial intelligence, and provides a model training method, a device and a storage medium, wherein the model training method comprises the following steps: obtaining a routing model to be trained; obtaining a plurality of sample transaction data and a plurality of attribute data; acquiring at least two target conditions; inputting a plurality of sample transaction data and a plurality of attribute data into a to-be-trained routing model to obtain at least two target statistical results; determining the number of statistical results which meet corresponding target conditions in at least two target statistical results, and if the number meets the model output conditions, determining the routing model to be trained as an alternative routing model; and if the quantity does not meet the model output condition, adjusting the parameters of the routing model to be trained, determining the routing model after the parameters are adjusted as the routing model to be trained until at least two statistical results meet the corresponding target conditions, and outputting one or more alternative routing models. By implementing the method and the device, the accuracy and the reasonability of the payment channel selection can be improved.
Description
Technical Field
The present application relates to the field of artificial intelligence technologies, and in particular, to a model training method, apparatus, and storage medium.
Background
In the payment system of the e-commerce platform, an important module is payment routing, and in a payment settlement link, a unique payment channel is found from a plurality of selectable payment channels according to transaction dimensions. But it is difficult to select an appropriate payment channel in the e-commerce platform. At present, when a model for selecting a payment channel is trained, model training is carried out by referring to single attribute data of the payment channel, for example, the reference cost is the minimum to be used as a training target until the model converges, and due to the fact that the reference factors are single in the training mode, the accuracy of the trained model is not high when the payment channel is selected.
Disclosure of Invention
Therefore, it is necessary to provide a model training method, device and storage medium for solving the above technical problems, so as to obtain a routing model for automatic selection of payment channels by integrating various attribute data of each payment channel, improve the accuracy of payment channel selection and optimize a payment channel selection scheme.
In a first aspect, the present application provides a model training method, the method comprising:
obtaining a routing model to be trained;
acquiring a plurality of sample transaction data in a training set and a plurality of attribute data of each payment channel in a plurality of payment channels associated with an e-commerce platform, wherein the plurality of attribute data comprise attribute data used for expressing the service quality of payment services provided by the payment channels for the e-commerce platform and/or attribute data used for expressing the payment cost paid by the e-commerce platform by using the payment channels;
acquiring at least two target conditions corresponding to at least two target attribute data in the plurality of attribute data, wherein one target attribute data corresponds to one target condition, and the target condition is used for defining a condition which needs to be met by a statistical result obtained by calculation based on the target attribute data when the plurality of sample transaction data are subjected to payment processing through the plurality of payment channels;
inputting the plurality of sample transaction data and the plurality of attribute data of each payment channel into the routing model to be trained, and obtaining a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme, wherein the at least two target statistical results have a corresponding relation with the at least two target conditions, one target statistical result corresponds to one target condition, and the target statistical result is statistical data obtained by calculation of the plurality of sample transaction data based on the target attribute data corresponding to the target condition when the payment channel selection scheme is used;
determining a first number of statistical results which meet corresponding target conditions in the at least two target statistical results, and if the first number meets model output conditions, determining the routing model to be trained as an alternative routing model; if the first quantity does not meet the model output condition, adjusting parameters of the to-be-trained routing model, determining the routing model after parameter adjustment as the to-be-trained routing model, outputting one or more alternative routing models until each target statistical result of the at least two target statistical results meets the corresponding target condition, wherein the one or more alternative routing models are used for determining the trained routing model, and the trained routing model is used for selecting a payment channel.
With reference to the first aspect, in some embodiments, the model output conditions include: the ratio of a first number of statistical results satisfying the corresponding target condition in the at least two target statistical results to the total number of statistical results in the at least two target statistical results is greater than a first preset threshold.
With reference to the first aspect, in some embodiments, after outputting one or more alternative routing models, the method further includes:
if the number of the output alternative routing models is one, determining the alternative routing models as trained routing models;
and if the number of the output alternative routing models is multiple, determining the trained routing model according to at least two target statistical results of each alternative routing model in the multiple alternative routing models.
With reference to the first aspect, in some embodiments, the determining a trained routing model according to at least two target statistics of each of the multiple candidate routing models includes:
acquiring the priorities of the at least two target attribute data corresponding to the at least two target statistical results, and determining a target statistical result corresponding to the target attribute data with the highest priority;
and comparing the target statistical result corresponding to the target attribute data with the highest priority of each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model with the optimal target statistical result corresponding to the target attribute data with the highest priority as the trained routing model.
With reference to the first aspect, in some embodiments, the determining a trained routing model according to at least two target statistics of each of the multiple candidate routing models includes:
obtaining a measurement unit of each target statistical result in the at least two target statistical results, and obtaining a unit conversion rule between the measurement units of each target statistical result;
for each alternative routing model in the multiple alternative routing models, converting the at least two target statistical results of the alternative routing model into a single statistical result corresponding to the alternative routing model according to the unit conversion rule, wherein the single statistical result comprises one statistical result;
and comparing the single statistical result corresponding to each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model corresponding to the statistical result with the optimal single statistical result as the trained routing model.
With reference to the first aspect, in some embodiments, the obtaining a plurality of pieces of sample transaction data comprises:
acquiring a historical transaction record associated with each payment channel in a plurality of payment channels associated with an e-commerce platform, wherein the historical transaction record comprises a transaction amount and a transaction time;
determining a plurality of historical transaction records associated with the plurality of payment channels as a plurality of sample transaction data.
With reference to the first aspect, in some embodiments, the plurality of attribute data includes a rate of a payment channel, a network stability of the payment channel, a maintenance time of the payment channel, and an arrival rate of the payment channel.
In a second aspect, the present application provides a model training apparatus, the apparatus comprising:
the first acquisition unit is used for acquiring a routing model to be trained;
the system comprises a first acquisition unit, a second acquisition unit and a control unit, wherein the first acquisition unit is used for acquiring a plurality of sample transaction data in a training set and a plurality of attribute data of each payment channel in a plurality of payment channels associated with an e-commerce platform, and the plurality of attribute data comprise attribute data used for expressing the service quality of payment services provided by the e-commerce platform by the payment channels and/or attribute data used for expressing the payment cost paid by the e-commerce platform by using the payment channels;
a third obtaining unit, configured to obtain at least two target conditions corresponding to at least two target attribute data in the plurality of attribute data, where one target attribute data corresponds to one target condition, and the target condition is used to define a condition that a statistical result calculated based on the target attribute data needs to be satisfied when the plurality of sample transaction data are subjected to payment processing through the plurality of payment channels;
the model training unit is used for inputting the plurality of sample transaction data and the plurality of attribute data of each payment channel into the routing model to be trained to obtain a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme, wherein the at least two target statistical results have a corresponding relation with the at least two target conditions, one target statistical result corresponds to one target condition, and the target statistical result is statistical data obtained by calculation of the plurality of sample transaction data based on the target attribute data corresponding to the target condition when the payment channel selection scheme is used;
the model training unit is further configured to determine a first number of statistical results satisfying corresponding target conditions among the at least two target statistical results, and if the first number satisfies a model output condition, determine the routing model to be trained as an alternative routing model; if the first quantity does not meet the model output condition, adjusting parameters of the to-be-trained routing model, determining the routing model after parameter adjustment as the to-be-trained routing model, outputting one or more alternative routing models until each target statistical result of the at least two target statistical results meets the corresponding target condition, wherein the one or more alternative routing models are used for determining the trained routing model, and the trained routing model is used for selecting a payment channel.
In combination with the second aspect, in some embodiments, the model output conditions include: the ratio of a first number of statistical results satisfying the corresponding target condition in the at least two target statistical results to the total number of statistical results in the at least two target statistical results is greater than a first preset threshold.
With reference to the second aspect, in some embodiments, the model training unit is further configured to determine the alternative routing model as a trained routing model if the number of the output alternative routing models is one;
and if the number of the output alternative routing models is multiple, determining the trained routing model according to at least two target statistical results of each alternative routing model in the multiple alternative routing models.
With reference to the second aspect, in some embodiments, the model training unit is specifically configured to obtain priorities of the at least two target attribute data corresponding to the at least two target statistical results, and determine a target statistical result corresponding to a target attribute data with a highest priority;
and comparing the target statistical result corresponding to the target attribute data with the highest priority of each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model with the optimal target statistical result corresponding to the target attribute data with the highest priority as the trained routing model.
With reference to the second aspect, in some embodiments, the model training unit is specifically configured to obtain a unit of measure of each of the at least two target statistical results, and obtain a unit conversion rule between the units of measure of each of the at least two target statistical results;
for each alternative routing model in the multiple alternative routing models, converting the at least two target statistical results of the alternative routing model into a single statistical result corresponding to the alternative routing model according to the unit conversion rule, wherein the single statistical result comprises one statistical result;
and comparing the single statistical result corresponding to each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model corresponding to the statistical result with the optimal single statistical result as the trained routing model.
With reference to the second aspect, in some embodiments, the second obtaining unit is specifically configured to obtain a historical transaction record associated with each payment channel of a plurality of payment channels associated with the e-commerce platform, where the historical transaction record includes a transaction amount and a transaction time;
determining a plurality of historical transaction records associated with the plurality of payment channels as a plurality of sample transaction data.
In combination with the second aspect, in some embodiments, the plurality of attribute data includes a rate of the payment channel, a network stability of the payment channel, a maintenance time of the payment channel, and an arrival rate of the payment channel.
In a third aspect, the present application provides a model training apparatus, including a processor, a memory, and a communication interface, where the processor, the memory, and the communication interface are connected to each other, where the communication interface is configured to receive and send data, the memory is configured to store program codes, and the processor is configured to call the program codes to perform a method as described in the first aspect and any possible implementation manner of the first aspect.
In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored therein; the computer program, when run on one or more processors, causes the terminal device to perform the method as described in the first aspect and any possible implementation form of the first aspect.
In the embodiment of the application, in the process of training a routing model, at least two target conditions corresponding to at least two target attribute data in a plurality of attribute data of a payment channel are set, further, after sample transaction data and the plurality of attribute data are input into a to-be-trained routing model, a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme are obtained, when the number of statistical results satisfying the corresponding target conditions in the at least two target statistical results satisfies a model output condition, an alternative routing model is output, and finally one or more alternative routing models are obtained, wherein the one or more alternative routing models are used for obtaining the trained routing model for selecting the payment channel. In other words, in the model training process, at least two target conditions corresponding to at least two target attribute data are considered comprehensively instead of a single training target, namely at least two target attribute data indexes are integrated to perform model training, so that the accuracy and the rationality of the payment channel selection are improved by the finally obtained trained routing model, and the payment channel selection scheme is optimized.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly described below.
Fig. 1 is a schematic flowchart of a model training method according to an embodiment of the present disclosure;
FIG. 2 is a schematic diagram of a model training process provided by an embodiment of the present application;
FIG. 3 is a schematic diagram of a model training apparatus according to an embodiment of the present disclosure;
fig. 4 is a schematic diagram of another model training apparatus according to an embodiment of the present application.
Detailed Description
The present invention is described in further detail below with reference to the attached drawing figures.
The terminology used in the following embodiments of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used in the specification of the present application and the appended claims, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
In this application, "at least one" means one or more, "a plurality" means two or more, "at least two" means two or three and three or more, "and/or" for describing an association relationship of associated objects, which means that there may be three relationships, for example, "a and/or B" may mean: only A, only B and both A and B are present, wherein A and B may be singular or plural. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. "at least one item(s) below" or similar expressions refer to any combination of these items. For example, at least one (one) of a, b, or c, may represent: a, b, c, "a and b," a and c, "" b and c, "or" a and b and c.
The embodiment of the application can acquire and process related data based on an artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human Intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result.
The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a robot technology, a biological recognition technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.
In order to more clearly describe the scheme of the present application, some drawings related to the present application are further described below.
Referring to fig. 1, fig. 1 is a schematic flow chart of a model training method according to an embodiment of the present disclosure. As shown in fig. 1, the method comprises the steps of:
step 120, obtaining a plurality of sample transaction data and a plurality of attribute data of each payment channel in a plurality of payment channels associated with an e-commerce platform, wherein the plurality of attribute data comprise attribute data used for representing service quality of payment services provided by the e-commerce platform by the payment channel and/or attribute data used for representing payment cost paid by the e-commerce platform by using the payment channel;
specifically, optionally, in the payment system of the e-commerce platform, an important module is a payment routing, and in a payment settlement link, a payment channel can be selected from a plurality of selectable payment channels to perform payment. Different e-commerce platforms contract with different payment channel providers, and one e-commerce platform contracts with multiple payment channel providers, which may refer to financial institutions providing payment functions, such as banks. The plurality of payment channels associated with the e-commerce platform in the present application may refer to each payment channel cooperating with the e-commerce platform, and if the payment channels are different, the values of the attribute data of the payment channels may also be different, for example, the attribute data is the rate of the payment channel, the rate value of the e-commerce platform and the payment channel 1 is 1%, and the rate value of the e-commerce platform and the payment channel 2 is 3%.
The application takes a plurality of attribute data corresponding to each payment channel as an example. The plurality of attribute data may include attribute data indicating quality of service of the payment channel for providing payment service for the e-commerce platform, such as network stability of the payment channel, a billing speed of the payment channel, and the like. Attribute data representing payment cost of the e-commerce platform for making payment by using the payment channel, such as the rate of the payment channel, can be further included in the plurality of attribute data.
Wherein the attribute data of the payment channel may be agreed by a contract made between the e-commerce platform and the payment channel provider, such as a rate of the payment channel. The attribute data of the payment channel may also be determined by the network characteristics of the payment channel provider itself, such as the network stability of the payment channel. The following exemplifies the attribute data of the payment channel:
channel rate: channel rates are one of the most important indicators for paying for routes. The general bank charges according to the amount, partly according to the transaction number, and the complicated point is the step charging, such as 10 in case of rate grade and 100 in case of rate grade.
Transaction limit: different payment channels may limit the amount cap for each transaction, as well as the amount limit for each account per day. Beyond this limit, payment channels need to be changed.
Channel stability, i.e., Quality of Service (QoS) of a channel: such as drop rate, network latency, and interface-capable Transaction Processing Systems (TPS).
Capital position: the capital position of the e-commerce company at the bank or third party platform. If the capital position is insufficient, the payment channel cannot be used for payment.
Channel account speed: for transfers, when funds go to the destination account.
Channel maintenance: when the payment channel is in maintenance, the payment channel needs to be selected according to the maintenance time of the payment channel, for example, the payment channel which is recently maintained is selected preferentially, and the payment channel which is to be maintained is selected later.
And flow distribution of each payment channel: the traffic distribution of each payment channel, also referred to as proportional distribution transaction of each payment channel, for example, the traffic distribution of each payment channel must be calculated in a ratio of 1:3: 4: a proportional distribution trade. Some payment channels are not good in all aspects, and as backup payment channels, in order to maintain long-term cooperation, some transaction amount needs to be distributed into the payment channels.
In this embodiment of the application, the sample transaction data used for training the routing model may be sample transaction data in a training set, or may be multiple historical transaction records associated with each payment channel. The sample transaction data may include information such as transaction amount, transaction time, transaction card number, etc. The plurality of historical transaction records may be transaction records occurring through the plurality of payment channels within a historical preset time period.
in one embodiment, the target condition may be a preset target of model training, and the at least two target conditions correspond to at least two target attribute data in the plurality of attribute data, and typically one target attribute data corresponds to one target condition. For example, the target condition may define a condition that needs to be satisfied by a statistical result calculated based on corresponding target attribute data when the plurality of pieces of sample transaction data are subjected to payment processing through the plurality of payment channels in the model training process. For example, the target condition may be that the overall cost charge is less than a certain threshold, which may be calculated based on the rate of the payment channel. As another example, the target condition may be that an average waiting time for the user transaction application is less than a certain threshold, and the average waiting time may be calculated based on network delays of respective payment channels. The target condition can also be that the account time period must be within a certain threshold value, the payment success rate is the lowest index value, and the like. The target condition may also define a priority of each attribute data, which may be used to indicate the importance of each attribute data, and the priority of each attribute data may be used to set a weight value of each attribute data in the routing model, such as setting a rate priority higher than a channel stability, and a rate weight that must be higher than a channel stability weight.
In the embodiment of the application, a plurality of sample transaction data and a plurality of attribute data of each payment channel are input into a to-be-trained routing model, so that a payment channel selection scheme is obtained, wherein the payment channel selection scheme comprises payment channels selected by the sample transaction data. Because at least two target conditions of model training are defined, correspondingly, the routing model to be trained outputs at least two target statistical results of the payment channel selection scheme, the number of the at least two target statistical results is the same as the number of the at least two defined target conditions, and a one-to-one correspondence relationship exists. The target statistical result is statistical data calculated based on target attribute data corresponding to target conditions when the plurality of sample transaction data use the payment channel selection scheme. Illustratively, the target condition defines that the total cost is less than a certain threshold, and the target statistical result corresponding to the target condition is the total cost calculated based on the rate of each payment channel when the payment channel selection scheme is used.
Further, after at least two target statistical results of the payment channel selection scheme are obtained, a first number of statistical results, which meet corresponding target conditions, of the at least two target statistical results is further determined, and if the first number meets model output conditions, the round of to-be-trained routing model is determined as an alternative routing model. The model output condition may be that the first number is greater than a certain threshold, or the model output condition may be that a ratio between the first number and the total number of the at least two target statistics is greater than a first preset threshold. Optionally, if the first number does not satisfy the model output condition, the parameter of the routing model to be trained needs to be adjusted, the routing model after the parameter adjustment is determined as the routing model to be trained, and the steps 110 to 150 are repeatedly executed, and the loop is iterated continuously until each target statistical result of the at least two target statistical results satisfies the corresponding target condition. Optionally, a reinforcement learning algorithm is used in the to-be-trained routing model, and a strategy iteration method is adopted for iterative training.
The model iterative training process of the present application is illustrated with reference to fig. 2, as shown in fig. 2, each attribute data of each payment channel, a plurality of historical transaction records, and at least two preset target conditions are input to a to-be-trained routing model for model training, in the iterative training process, in each iteration, the to-be-trained routing model in the iteration is obtained, and at least two target statistical results when the to-be-trained routing model is used to select the payment channel for the plurality of historical transaction records are calculated, where the at least two target statistical results may include, but are not limited to, total cost, payment success rate, and the like. The at least two target statistics may be compared with the corresponding target conditions, respectively, to determine whether the target statistics satisfy the corresponding target conditions. And further, determining whether the at least two target statistical results output the model output condition according to the comparison result. Optionally, when the target statistical result greater than or equal to a certain proportion threshold satisfies the corresponding target condition, the round of to-be-trained routing model may be determined as the alternative routing model, and for example, if the proportion threshold is 70%, if more than 70% of the target statistical results satisfy the corresponding target condition, the round of to-be-trained routing model is determined as the alternative routing model. For example, the to-be-trained routing model performs payment channel selection calculation on 100 pieces of sample transaction data to obtain 10 target statistics, wherein 7 target statistics satisfy corresponding target conditions, and then the to-be-trained routing model can be determined as an alternative routing model.
If at least two target statistical results output by the iterative routing model to be trained do not meet preset model output conditions, the iterative training is further required to be continuously performed, parameters of the routing model to be trained are readjusted, and the training is continued, which may be to adjust weight values of each attribute data in the routing model to be trained. And continuously circulating the process, and if at least two target statistical results obtained by calculation of the route model to be trained of a certain iteration meet the model output condition, determining the route model to be trained of the iteration as an alternative route model, so as to obtain at least one alternative route model. And further manually checking whether each alternative routing model meets the expectation, if not, indicating that the target condition is unreasonable, readjusting the target condition, and if so, determining the trained routing model from the at least one alternative routing model.
In the embodiment of the present application, if only one alternative routing model is obtained, the alternative routing model may be optimal, and is determined as the trained routing model. If the number of the candidate routing models is greater than the preset threshold, for example, greater than 100, it indicates that the preset target condition is unreasonable, the target condition may be reset, and the iterative training of the routing model may be performed again. If the number of the trained AI routing models is less than the set threshold and greater than 1, one routing model may be selected from the plurality of candidate routing models as the trained routing model. The selection mode for selecting one routing model from the multiple candidate routing models may be set according to actual needs, and may be selected according to at least two target statistical results of each candidate routing model in the multiple candidate routing models. Two alternative options are provided below for illustration:
in the first mode, at least two target statistical results corresponding to each alternative routing model are obtained, priorities of target attribute data corresponding to the at least two target statistical results are obtained, and a target statistical result corresponding to the target attribute data with the highest priority is determined, for example, if the rate is the highest priority, the cost fee is the target statistical result corresponding to the highest priority. And comparing the data size of the target statistical result corresponding to the target attribute data with the highest priority in each alternative routing model, thereby determining the trained routing model. For example, if the cost is the target statistical result corresponding to the highest priority, the candidate routing model with the lowest cost in the candidate routing models is determined as the trained routing model.
In the second mode, although the measurement units of the target statistical results may be different, unit conversion rules among the measurement units may be preset, and the target statistical results are converted into the statistical results of the same measurement unit according to the unit conversion rules, so that the optimal alternative routing model is determined as the trained routing model. For example, if the unit of measure of the cost is element and the unit of measure of the charge arrival rate is minute, the unit conversion rule between the cost and the charge arrival rate may be: the 1 yuan and 10 minutes may be converted to each other, that is, 10 minutes may be converted to 1 yuan, or 1 yuan may be converted to 10 minutes.
In some optional manners, the plurality of candidate routing models and at least two target statistical results corresponding to each candidate routing model may also be output, and a human decision may select which one is used as the trained routing model.
After the trained routing model is determined, when transaction data is received subsequently, the transaction data is input into the trained routing model, and the trained routing model can output a payment channel which is selected by the transaction data.
In the embodiment of the application, in the process of training a routing model, at least two target conditions corresponding to at least two target attribute data in a plurality of attribute data of a payment channel are set, further, after sample transaction data and the plurality of attribute data are input into a to-be-trained routing model, a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme are obtained, when the number of statistical results satisfying the corresponding target conditions in the at least two target statistical results satisfies a model output condition, an alternative routing model is output, and finally one or more alternative routing models are obtained, wherein the one or more alternative routing models are used for obtaining the trained routing model for selecting the payment channel. In other words, in the model training process, at least two target conditions corresponding to at least two target attribute data are considered comprehensively instead of a single training target, namely at least two target attribute data indexes are integrated to perform model training, so that the accuracy and the rationality of the payment channel selection are improved by the finally obtained trained routing model, and the payment channel selection scheme is optimized.
Please refer to fig. 3, which provides a schematic structural diagram of a model training apparatus according to an embodiment of the present application. As shown in fig. 3, the model training apparatus may include:
a first obtaining unit 10, configured to obtain a routing model to be trained;
a second obtaining unit 11, configured to obtain a plurality of sample transaction data in a training set and a plurality of attribute data of each payment channel of a plurality of payment channels associated with an e-commerce platform, where the plurality of attribute data includes attribute data used for indicating quality of service of a payment service provided by the payment channel for the e-commerce platform and/or attribute data used for indicating a payment cost for the e-commerce platform to pay using the payment channel;
a third obtaining unit 12, configured to obtain at least two target conditions corresponding to at least two target attribute data in the plurality of attribute data, where one target attribute data corresponds to one target condition, and the target condition is used to define a condition that a statistical result calculated based on the target attribute data needs to be satisfied when the plurality of sample transaction data are subjected to payment processing through the plurality of payment channels;
a model training unit 13, configured to input the multiple pieces of sample transaction data and the multiple pieces of attribute data of each payment channel into the to-be-trained routing model, so as to obtain a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme, where the at least two target statistical results have a correspondence with the at least two target conditions, and one target statistical result corresponds to one target condition, where the target statistical result is statistical data calculated based on target attribute data corresponding to the target condition when the multiple pieces of sample transaction data use the payment channel selection scheme;
the model training unit 13 is further configured to determine a first number of statistical results satisfying corresponding target conditions among the at least two target statistical results, and if the first number satisfies a model output condition, determine the routing model to be trained as an alternative routing model; if the first quantity does not meet the model output condition, adjusting parameters of the routing model to be trained, determining the routing model after parameter adjustment as the routing model to be trained, and outputting one or more alternative routing models until each statistical result of the at least two statistical results meets the corresponding target condition, wherein the one or more alternative routing models are used for determining the trained routing model, and the trained routing model is used for selecting a payment channel.
In one possible design, the model output conditions include: the ratio of a first number of statistical results satisfying the corresponding target condition in the at least two target statistical results to the total number of statistical results in the at least two target statistical results is greater than a first preset threshold.
In a possible design, the model training unit 13 is further configured to determine the alternative routing model as a trained routing model if the number of the output alternative routing models is one;
and if the number of the output alternative routing models is multiple, determining the trained routing model according to at least two target statistical results of each alternative routing model in the multiple alternative routing models.
In a possible design, the model training unit 13 is specifically configured to obtain priorities of the at least two target attribute data corresponding to the at least two target statistical results, and determine a target statistical result corresponding to a target attribute data with a highest priority;
and comparing the target statistical result corresponding to the target attribute data with the highest priority of each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model with the optimal target statistical result corresponding to the target attribute data with the highest priority as the trained routing model.
In a possible design, the model training unit 13 is specifically configured to obtain a measurement unit of each of the at least two target statistical results, and obtain a unit transformation rule between the measurement units of each of the at least two target statistical results;
for each alternative routing model in the multiple alternative routing models, converting the at least two target statistical results of the alternative routing model into a single statistical result corresponding to the alternative routing model according to the unit conversion rule, wherein the single statistical result comprises one statistical result;
and comparing the single statistical result corresponding to each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model corresponding to the statistical result with the optimal single statistical result as the trained routing model.
In a possible design, the second obtaining unit 11 is specifically configured to obtain a historical transaction record associated with each payment channel of a plurality of payment channels associated with the e-commerce platform, where the historical transaction record includes a transaction amount and a transaction time;
determining a plurality of historical transaction records associated with the plurality of payment channels as a plurality of sample transaction data.
In one possible design, the plurality of attribute data includes a rate of the payment channel, a network stability of the payment channel, a maintenance time of the payment channel, and an arrival rate of the payment channel.
For a specific description of the embodiment of the apparatus shown in fig. 3, reference may be made to the foregoing specific description of the embodiment of the method shown in fig. 1, which is not repeated herein.
Referring to fig. 4, which is a schematic structural diagram of another model training device provided in the present embodiment, as shown in fig. 4, the model training device 1000 may include: at least one processor 1001, such as a CPU, at least one communication interface 1003, memory 1004, at least one communication bus 1002. Wherein a communication bus 1002 is used to enable connective communication between these components. The communication interface 1003 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1004 may be a high-speed RAM memory or a non-volatile memory (e.g., at least one disk memory). The memory 1004 may optionally be at least one storage device located remotely from the processor 1001. As shown in fig. 4, memory 1004, which is a type of computer storage medium, may include an operating system, network communication units, and program instructions.
In the model training apparatus 1000 shown in fig. 4, the processor 1001 may be configured to load program instructions stored in the memory 1004 and specifically perform the following operations:
obtaining a routing model to be trained;
acquiring a plurality of sample transaction data and a plurality of attribute data of each payment channel in a plurality of payment channels associated with an e-commerce platform, wherein the attribute data comprises attribute data used for representing the service quality of payment services provided by the payment channels for the e-commerce platform and/or attribute data used for representing the payment cost of payment carried out by the e-commerce platform by using the payment channels;
acquiring at least two target conditions corresponding to at least two target attribute data in the plurality of attribute data, wherein one target attribute data corresponds to one target condition, and the target condition is used for defining a condition which needs to be met by a statistical result obtained by calculation based on the target attribute data when the plurality of sample transaction data are subjected to payment processing through the plurality of payment channels;
inputting the plurality of sample transaction data and the plurality of attribute data of each payment channel into the routing model to be trained, and obtaining a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme, wherein the payment channel selection scheme comprises the payment channel selected by each sample transaction data in the plurality of sample transaction data, the at least two target statistical results and the at least two target conditions have a corresponding relationship, one target statistical result corresponds to one target condition, and the target statistical result refers to statistical data obtained by calculation of the plurality of sample transaction data based on the target attribute data corresponding to the target condition when the payment channel selection scheme is used;
determining a first number of statistical results which meet corresponding target conditions in the at least two target statistical results, and if the first number meets model output conditions, determining the routing model to be trained as an alternative routing model; if the first quantity does not meet the model output condition, adjusting parameters of the routing model to be trained, determining the routing model after parameter adjustment as the routing model to be trained, and outputting one or more alternative routing models until each statistical result of the at least two statistical results meets the corresponding target condition, wherein the one or more alternative routing models are used for determining the trained routing model, and the trained routing model is used for selecting a payment channel.
It should be noted that, for a specific implementation process, reference may be made to specific descriptions of the method embodiment shown in fig. 1, which are not described herein again.
An embodiment of the present application further provides a computer storage medium, where the computer storage medium may store a plurality of instructions, where the instructions are suitable for being loaded by a processor and executing the method steps in the embodiment shown in fig. 1, and a specific execution process may refer to a specific description of the embodiment shown in fig. 1, which is not described herein again.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The procedures or functions described in accordance with the embodiments of the present application occur, in whole or in part, when the computer program instructions are loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (e.g., coaxial cable, fiber optic, digital subscriber line) or wirelessly (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The available media may be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid state drives), among others.
One of ordinary skill in the art will appreciate that all or part of the processes in the methods of the above embodiments may be implemented by hardware related to instructions of a computer program, which may be stored in a computer-readable storage medium, and when executed, may include the processes of the above method embodiments. And the aforementioned storage medium includes: various media capable of storing program codes, such as ROM or RAM, magnetic or optical disks, etc.
Claims (10)
1. A method of model training, comprising:
obtaining a routing model to be trained;
obtaining a plurality of sample transaction data and a plurality of attribute data of each payment channel in a plurality of payment channels associated with an e-commerce platform;
acquiring at least two target conditions corresponding to at least two target attribute data in the plurality of attribute data, wherein the target conditions are used for defining conditions which need to be met by statistical results obtained by calculation based on the target attribute data corresponding to the target conditions when the plurality of sample transaction data are subjected to payment processing through the plurality of payment channels;
inputting the plurality of sample transaction data and the plurality of attribute data of each payment channel into the routing model to be trained, and obtaining a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme, wherein the at least two target statistical results have a corresponding relation with the at least two target conditions, and the target statistical results are statistical data obtained by calculation of the plurality of sample transaction data based on the target attribute data corresponding to the target conditions when the payment channel selection scheme is used;
determining a first number of statistical results which meet corresponding target conditions in the at least two target statistical results, and if the first number meets model output conditions, determining the routing model to be trained as an alternative routing model; if the first quantity does not meet the model output condition, adjusting parameters of the to-be-trained routing model, determining the routing model after parameter adjustment as the to-be-trained routing model, outputting one or more alternative routing models until each target statistical result of the at least two target statistical results meets the corresponding target condition, wherein the one or more alternative routing models are used for determining the trained routing model, and the trained routing model is used for selecting a payment channel.
2. The method of claim 1, wherein the model output conditions comprise: the ratio of a first number of statistical results satisfying the corresponding target condition in the at least two target statistical results to the total number of statistical results in the at least two target statistical results is greater than a first preset threshold.
3. The method of claim 1 or 2, wherein after outputting the one or more alternative routing models, further comprising:
if the number of the output alternative routing models is one, determining the alternative routing models as trained routing models;
and if the number of the output alternative routing models is multiple, determining the trained routing model according to at least two target statistical results of each alternative routing model in the multiple alternative routing models.
4. The method of claim 3, wherein determining the trained routing model based on the at least two target statistics for each of the plurality of alternative routing models comprises:
acquiring the priorities of the at least two target attribute data corresponding to the at least two target statistical results, and determining a target statistical result corresponding to the target attribute data with the highest priority;
and comparing the target statistical result corresponding to the target attribute data with the highest priority of each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model with the optimal target statistical result corresponding to the target attribute data with the highest priority as the trained routing model.
5. The method of claim 3, wherein determining the trained routing model based on the at least two target statistics for each of the plurality of alternative routing models comprises:
obtaining a measurement unit of each target statistical result in the at least two target statistical results, and obtaining a unit conversion rule between the measurement units of each target statistical result;
for each alternative routing model in the multiple alternative routing models, converting the at least two target statistical results of the alternative routing model into a single statistical result corresponding to the alternative routing model according to the unit conversion rule, wherein the single statistical result comprises one statistical result;
and comparing the single statistical result corresponding to each alternative routing model in the plurality of alternative routing models, and determining the alternative routing model corresponding to the statistical result with the optimal single statistical result as the trained routing model.
6. The method of claim 1, wherein said obtaining a plurality of sample transaction data comprises:
acquiring a historical transaction record associated with each payment channel in a plurality of payment channels associated with an e-commerce platform, wherein the historical transaction record comprises a transaction amount and a transaction time;
determining a plurality of historical transaction records associated with the plurality of payment channels as a plurality of sample transaction data.
7. The method of claim 1, wherein the plurality of attribute data includes a rate for a payment channel, a network stability for a payment channel, a maintenance time for a payment channel, and a billing rate for a payment channel.
8. A model training apparatus, comprising:
the first acquisition unit is used for acquiring a routing model to be trained;
the second acquisition unit is used for acquiring a plurality of sample transaction data in a training set and a plurality of attribute data of each payment channel in a plurality of payment channels associated with the e-commerce platform;
a third obtaining unit, configured to obtain at least two target conditions corresponding to at least two target attribute data in the plurality of attribute data, where the target conditions are used to define conditions that need to be satisfied by a statistical result calculated based on the target attribute data corresponding to the target conditions when the plurality of sample transaction data are subjected to payment processing through the plurality of payment channels;
the model training unit is used for inputting the plurality of sample transaction data and the plurality of attribute data of each payment channel into the routing model to be trained to obtain a payment channel selection scheme and at least two target statistical results of the payment channel selection scheme, wherein the at least two target statistical results have a corresponding relation with at least two target conditions, and the target statistical results are statistical data obtained by calculation of the plurality of sample transaction data based on the target attribute data corresponding to the target conditions when the payment channel selection scheme is used;
the model training unit is further configured to determine a first number of statistical results satisfying corresponding target conditions among the at least two target statistical results, and if the first number satisfies a model output condition, determine the routing model to be trained as an alternative routing model; if the first quantity does not meet the model output condition, adjusting parameters of the to-be-trained routing model, determining the routing model after parameter adjustment as the to-be-trained routing model, outputting one or more alternative routing models until each target statistical result of the at least two target statistical results meets the corresponding target condition, wherein the one or more alternative routing models are used for determining the trained routing model, and the trained routing model is used for selecting a payment channel.
9. A model training apparatus comprising a processor, a memory and a communication interface, the processor, the memory and the communication interface being interconnected, wherein the communication interface is configured to receive and transmit data, the memory is configured to store program code, and the processor is configured to invoke the program code to perform the method of any of claims 1 to 7.
10. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium; the computer program, when run on one or more processors, performs the method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110913896.2A CN113627900A (en) | 2021-08-10 | 2021-08-10 | Model training method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110913896.2A CN113627900A (en) | 2021-08-10 | 2021-08-10 | Model training method, device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113627900A true CN113627900A (en) | 2021-11-09 |
Family
ID=78383952
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110913896.2A Pending CN113627900A (en) | 2021-08-10 | 2021-08-10 | Model training method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113627900A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116109395A (en) * | 2023-03-27 | 2023-05-12 | 北京沐融信息科技股份有限公司 | Intelligent route updating method, device, electronic equipment and medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109345302A (en) * | 2018-09-27 | 2019-02-15 | 腾讯科技（深圳）有限公司 | Machine learning model training method, device, storage medium and computer equipment |
| WO2019196261A1 (en) * | 2018-04-08 | 2019-10-17 | 平安科技（深圳）有限公司 | Method for pushing protocol file, and terminal device |
| CN110458668A (en) * | 2019-08-15 | 2019-11-15 | 中国银行股份有限公司 | Determine the method and device of Products Show algorithm |
| CN110532467A (en) * | 2019-08-22 | 2019-12-03 | 上海易点时空网络有限公司 | Activity recommendation method and device, equipment, storage medium based on push model |
| CN111143685A (en) * | 2019-12-30 | 2020-05-12 | 第四范式（北京）技术有限公司 | Recommendation system construction method and device |
| CN112435021A (en) * | 2020-11-25 | 2021-03-02 | 平安普惠企业管理有限公司 | Transaction request processing method, device, equipment and medium based on artificial intelligence |
| CN112508301A (en) * | 2020-12-21 | 2021-03-16 | 北京梧桐车联科技有限责任公司 | Method and device for predicting charging load of electric vehicle, and storage medium |
-
2021
- 2021-08-10 CN CN202110913896.2A patent/CN113627900A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019196261A1 (en) * | 2018-04-08 | 2019-10-17 | 平安科技（深圳）有限公司 | Method for pushing protocol file, and terminal device |
| CN109345302A (en) * | 2018-09-27 | 2019-02-15 | 腾讯科技（深圳）有限公司 | Machine learning model training method, device, storage medium and computer equipment |
| CN110458668A (en) * | 2019-08-15 | 2019-11-15 | 中国银行股份有限公司 | Determine the method and device of Products Show algorithm |
| CN110532467A (en) * | 2019-08-22 | 2019-12-03 | 上海易点时空网络有限公司 | Activity recommendation method and device, equipment, storage medium based on push model |
| CN111143685A (en) * | 2019-12-30 | 2020-05-12 | 第四范式（北京）技术有限公司 | Recommendation system construction method and device |
| CN112435021A (en) * | 2020-11-25 | 2021-03-02 | 平安普惠企业管理有限公司 | Transaction request processing method, device, equipment and medium based on artificial intelligence |
| CN112508301A (en) * | 2020-12-21 | 2021-03-16 | 北京梧桐车联科技有限责任公司 | Method and device for predicting charging load of electric vehicle, and storage medium |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116109395A (en) * | 2023-03-27 | 2023-05-12 | 北京沐融信息科技股份有限公司 | Intelligent route updating method, device, electronic equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100751965B1 (en) | method and system for predicting attrition customers | |
| CN109672795A (en) | Call center resource management method and device, electronic equipment, storage medium | |
| CN108733508A (en) | Method and system for controlling data backup | |
| CN101018164A (en) | A TCP/IP network performance evaluation prediction method | |
| CN111435463A (en) | Data processing method and related equipment and system | |
| CN110838031A (en) | Data operation method and device based on ABtest | |
| CN106454437A (en) | Streaming media service rate prediction method and device | |
| JP6356901B2 (en) | Adaptive connection system based on flexible risk transfer structure and method for the system | |
| CN109886699A (en) | Activity recognition method and device, electronic equipment, storage medium | |
| CN112001556A (en) | Reservoir downstream water level prediction method based on deep learning model | |
| CN107274215A (en) | Flight prices Forecasting Methodology, device, equipment and storage medium | |
| CN113627900A (en) | Model training method, device and storage medium | |
| CN110058937A (en) | For dispatching the method, equipment and computer program product of dedicated processes resource | |
| WO2023124363A1 (en) | Inventory determination method and apparatus | |
| CN109492863A (en) | The automatic generation method and device of financial document | |
| CN113302644B (en) | Transaction plan management system | |
| CN110009154B (en) | Refund prediction method and device, terminal equipment and storage medium | |
| JP5831363B2 (en) | Manufacturing lead time prediction device, manufacturing lead time prediction method, and computer program | |
| CN111510473B (en) | Access request processing method and device, electronic equipment and computer readable medium | |
| WO2018182179A1 (en) | Method and apparatus for asset management | |
| US11587092B2 (en) | System for dynamic prediction of interchange rates for credit card transaction processing | |
| CN112907362A (en) | Loan transaction processing method and device, electronic equipment and storage medium | |
| CN114756774A (en) | Travel scheme recommendation method, travel scheme recommendation device, model training equipment and storage medium | |
| CN113869596A (en) | Task prediction processing method, device, product and medium | |
| CN113570114A (en) | Intelligent matching method and system for resource service and computer equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |