Disclosure of Invention
The application provides a user determining method and device based on financial products and electronic equipment, and aims to at least solve the technical problem that determining efficiency is low when a user with purchasing intention for the financial products is determined in the prior art.
According to one aspect of the present application, there is provided a financial product-based user determination method, comprising: inputting product information of a target financial product into a target model to obtain a user portrait to be predicted corresponding to the target financial product, wherein the target model is a neural network model obtained by training according to reference user portraits corresponding to each reference user in L reference users, each reference user portraits comprises N user attribute information corresponding to the reference user, the user portrait to be predicted comprises N user attribute information corresponding to the user to be predicted, the reference user is a user who purchases any financial product, the user to be predicted is a user with purchase intention for the target financial product, and L and N are positive integers; obtaining M alternative user portraits corresponding to M alternative users, wherein M is a positive integer, each alternative user portrait in the M alternative user portraits comprises N user attribute information corresponding to one alternative user in the M alternative users, and the alternative users are users who do not purchase target gold melt products; detecting the similarity between the user portrait to be predicted and each of the M candidate user portraits; and determining at least one target user from the M candidate users according to the similarity corresponding to each candidate user image, and sending the product information of the target financial product to the terminal equipment of the target user.
Optionally, the financial product-based user determination method further comprises: extracting features of reference user images corresponding to each reference user in the L reference users to obtain N first feature vectors corresponding to the reference user, wherein each first feature vector corresponds to user attribute information of one reference user; extracting the characteristics of the product information of the financial products purchased by each reference user to obtain a second characteristic vector corresponding to the reference user, wherein the second characteristic vector corresponding to each reference user is used for representing the product types corresponding to the financial products purchased by the reference user; and generating a target model according to the N first feature vectors and the N second feature vectors.
Optionally, the financial product-based user determination method further comprises: clustering N first feature vectors to obtain P vector sets, wherein each vector set comprises the same Q first feature vectors, and P and Q are positive integers; selecting a first feature vector from each vector set as a first target feature vector corresponding to the vector set to obtain P first target feature vectors; determining S first mapping relations according to the P first target feature vectors and the second feature vectors corresponding to each of the L reference users, wherein S is a positive integer, and each first mapping relation is used for representing the mapping relation between one first target feature vector of the P first target feature vectors and the second feature vector corresponding to one of the L reference users; splicing a first target feature vector and a second feature vector corresponding to each first mapping relation in the S first mapping relations to obtain a third feature vector corresponding to the first mapping relation; and generating a target model according to the third feature vector corresponding to each first mapping relation in the S first mapping relations.
Optionally, the financial product-based user determination method further comprises: inputting a third feature vector corresponding to each first mapping relation in the S first mapping relations into a generated countermeasure network, wherein the generated countermeasure network comprises a generator and a discriminator, the generator is used for generating user attribute information corresponding to the virtual user according to the third feature vector, and the discriminator is used for determining the similarity of the virtual user and the real user according to the user attribute information generated by the generator; performing multiple countermeasure training on the generator and the discriminator according to the similarity between the virtual user and the real user until the similarity between the virtual user and the real user determined by the target discriminator according to the user attribute information generated by the target generator is greater than a preset threshold, wherein the target discriminator is a generator after multiple countermeasure training, and the target generator is a generator after multiple countermeasure training; a target model is generated based on the target generator and the target arbiter.
Optionally, the financial product-based user determination method further comprises: determining a financial tag corresponding to the target financial product according to the product information of the target financial product, wherein the financial tag is used for representing the product type corresponding to the target financial product; inputting the financial tag into each target sub-generator in N target sub-generators to obtain user attribute information of a user to be predicted corresponding to each target sub-generator, wherein the target generator comprises N target sub-generators, and the target sub-generators are used for predicting one type of user attribute information; and determining the user portrait to be predicted according to the user attribute information of the user to be predicted corresponding to each target sub-generator in the N target sub-generators.
Optionally, the N pieces of user attribute information in the financial product-based user determination method at least include region attribute information, behavior attribute information and risk attribute information, where the region attribute information is used to represent region information of a user, the behavior attribute information is used to represent consumption behavior information of the user, and the risk attribute information is used to represent credit risk information of the user.
Optionally, the financial product-based user determination method further comprises: n sub-similarities between the candidate user portrait of each candidate user and the user portrait to be predicted are determined according to N user attribute information corresponding to each candidate user and N user attribute information corresponding to the user to be predicted, wherein each sub-similarity is used for representing the similarity between one user attribute information corresponding to each candidate user and one user attribute information corresponding to the user to be predicted, and two user attribute information corresponding to each sub-similarity belong to the same type of attribute information; and determining the similarity between the candidate user portrait and the user portrait to be predicted according to N sub-similarities between the candidate user portrait and the user portrait to be predicted of each candidate user.
According to another aspect of the present application, there is also provided a financial product-based user determination apparatus including: the input unit is used for inputting product information of a target financial product into a target model to obtain a user portrait to be predicted corresponding to the target financial product, wherein the target model is a neural network model obtained by training according to reference user portraits corresponding to each reference user in L reference users, each reference user portraits comprises N user attribute information corresponding to the reference user, the user portrait to be predicted comprises N user attribute information corresponding to the user to be predicted, the reference user is a user who purchases any financial product, the user to be predicted is a user who has purchase intention for the target financial product, and L and N are positive integers; the acquisition unit is used for acquiring M alternative user portraits corresponding to M alternative users, wherein M is a positive integer, each alternative user portrait in the M alternative user portraits comprises N user attribute information corresponding to one alternative user in the M alternative users, and the alternative users are users who do not purchase the target metal fusion product; the detection unit is used for detecting the similarity between the user portrait to be predicted and each candidate user portrait in the M candidate user portraits; and the determining unit is used for determining at least one target user from the M candidate users according to the similarity corresponding to each candidate user image and sending the product information of the target financial product to the terminal equipment of the target user.
According to another aspect of the present application, there is also provided a computer readable storage medium having a computer program stored therein, wherein the computer readable storage medium is controlled to perform any one of the above-mentioned financial product-based user determination methods when the computer program is run.
According to another aspect of the present application, there is also provided an electronic device, wherein the electronic device comprises one or more processors and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the financial product based user determination method of any of the above.
In the method, product information of a target financial product is firstly input into a target model to obtain a user portrait to be predicted corresponding to the target financial product, wherein the target model is a neural network model obtained by training according to reference user images corresponding to each of L reference users, each reference user image comprises N user attribute information corresponding to the reference user, the user portrait to be predicted comprises N user attribute information corresponding to the user to be predicted, the reference user is a user who purchases any financial product, the user to be predicted is a user who has purchase intention for the target financial product, L and N are positive integers, then M alternative user portraits corresponding to M alternative users are obtained, M is a positive integer, each alternative user portraits in the M alternative user images comprises N user attribute information corresponding to one alternative user in the M alternative users, each alternative user is a user who does not purchase the target financial product, then the degree between the user portrait to be predicted and each alternative user image in the M alternative user images is detected, and the degree of similarity between the target financial product and at least one target financial product is determined according to the terminal similarity of each alternative user image.
From the above, it can be seen that, the target model in the present application is obtained by training according to the reference user portrait corresponding to each reference user in the L reference users, and the reference user portrait includes N user attribute information corresponding to the reference user, so that it can be deduced that the user portrait to be predicted in the present application is obtained by training the plurality of user attribute information included in the reference user portrait corresponding to each reference user, and compared with the manner of determining the user corresponding to the financial product based on a single user consumption feature in the prior art, the present application does not determine the target user corresponding to the financial product according to a single feature of the candidate user, but determines the target user corresponding to the financial product by judging the similarity between the candidate user portrait and the user portrait to be predicted, thereby achieving the purpose of improving the accuracy when determining the user having the purchase intention for the target financial product.
Therefore, the technical scheme of the application utilizes the target model to generate the user portrait to be predicted, and determines the target user corresponding to the target financial product according to the similarity between the user portrait to be predicted and each candidate user, thereby realizing the purpose of improving the accuracy of determining the user with the purchase intention on the target financial product, further realizing the technical effect of improving the determination efficiency of determining the user with the purchase intention on the target financial product, and further solving the technical problem of low determination efficiency when determining the user with the purchase intention on the financial product in the prior art.
Detailed Description
In order to make the present application solution better understood by those skilled in the art, the following description will be made in detail and with reference to the accompanying drawings in the embodiments of the present application, it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, shall fall within the scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that embodiments of the present application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be further noted that, the relevant information (including, but not limited to, product information and user attribute information of the target financial product) and the data (including, but not limited to, data for presentation and data analyzed) related to the present application are information and data authorized by the user or sufficiently authorized by each party. For example, an interface is provided between the system and the relevant user or institution, before acquiring the relevant information, the system needs to send an acquisition request to the user or institution through the interface, and acquire the relevant information after receiving the consent information fed back by the user or institution.
The present application is further illustrated below in conjunction with various embodiments.
Example 1
In accordance with embodiments of the present application, there is provided an embodiment of a financial product-based user determination method, it being noted that the steps illustrated in the flowchart of the figures may be performed in a computer system, such as a set of computer-executable instructions, and, although a logical order is illustrated in the flowchart, in some cases, the steps illustrated or described may be performed in an order other than that illustrated herein.
The present application provides a financial product-based user determination system (abbreviated as determination system) for performing a financial product-based user determination method in the present application, and fig. 1 is a flowchart of an alternative financial product-based user determination method according to an embodiment of the present application, as shown in fig. 1, and the method includes the steps of:
and step S101, inputting product information of the target financial product into the target model to obtain a user portrait to be predicted corresponding to the target financial product.
In step S101, the target model is a neural network model obtained by training according to a reference user image corresponding to each reference user of the L reference users, each reference user image includes N user attribute information corresponding to the reference user, the user portrait to be predicted includes N user attribute information corresponding to the user to be predicted, the reference user is a user who purchases any financial product, the user to be predicted is a user who has a purchase intention for the target financial product, and L and N are both positive integers.
Optionally, the target model is an antagonistic neural network model, and the target model includes a target generator and a target arbiter, and the N pieces of user attribute information corresponding to the reference user include, but are not limited to, region attribute information, behavior attribute information, and risk attribute information of the reference user.
Step S102, M alternative user portraits corresponding to the M alternative users are obtained.
In step S102, M is a positive integer, and each of the M candidate user images includes N user attribute information corresponding to one of the M candidate users, where the candidate users are users who have not purchased the target metal-fusion product.
Optionally, the N pieces of user attribute information corresponding to the candidate user include, but are not limited to, regional attribute information, behavioral attribute information, and risk attribute information of the candidate user.
Step S103, detecting the similarity between the user portrait to be predicted and each of the M candidate user portraits.
Optionally, the determining system determines N sub-similarities between each candidate user and the user to be predicted, where each sub-similarity characterizes a similarity between one user attribute information corresponding to the candidate user and one user attribute information corresponding to the user to be predicted, and two user attribute information corresponding to each sub-similarity belong to the same type of attribute information, and then performs weighted summation on N sub-similarities between the candidate user portrait of each candidate user and the user portrait to be predicted to obtain the similarity between the candidate user portrait and the user portrait to be predicted.
Step S104, determining at least one target user from M candidate users according to the similarity corresponding to each candidate user image, and sending product information of the target financial product to terminal equipment of the target user.
Optionally, the determining system may further sort the M candidate users according to the similarity corresponding to each candidate user image, and use the candidate users in the sorting result that are greater than or equal to the preset similarity threshold as the target users.
From the above, it can be seen that, the target model in the present application is obtained by training according to the reference user portrait corresponding to each reference user in the L reference users, and the reference user portrait includes N user attribute information corresponding to the reference user, so that it can be deduced that the user portrait to be predicted in the present application is obtained by training the plurality of user attribute information included in the reference user portrait corresponding to each reference user, and compared with the manner of determining the user corresponding to the financial product based on a single user consumption feature in the prior art, the present application does not determine the target user corresponding to the financial product according to a single feature of the candidate user, but determines the target user corresponding to the financial product by judging the similarity between the candidate user portrait and the user portrait to be predicted, thereby achieving the purpose of improving the accuracy when determining the user having the purchase intention for the target financial product.
Therefore, the technical scheme of the application utilizes the target model to generate the user portrait to be predicted, and determines the target user corresponding to the target financial product according to the similarity between the user portrait to be predicted and each candidate user, thereby realizing the purpose of improving the accuracy of determining the user with the purchase intention on the target financial product, further realizing the technical effect of improving the determination efficiency of determining the user with the purchase intention on the target financial product, and further solving the technical problem of low determination efficiency when determining the user with the purchase intention on the financial product in the prior art.
In an alternative embodiment, the target model in the financial product based user determination method is trained by the steps of: firstly, determining that a system performs feature extraction on reference user images corresponding to each reference user in L reference users to obtain N first feature vectors corresponding to the reference user, wherein each first feature vector corresponds to user attribute information of one reference user, then performing feature extraction on product information of financial products purchased by each reference user to obtain a second feature vector corresponding to the reference user, wherein the second feature vector corresponding to each reference user is used for representing product types corresponding to the financial products purchased by the reference user, and then generating a target model according to the N first feature vectors and the second feature vectors.
Optionally, the set u= { U1 ,U2 ,……,UL U, wherei Representing the representation of the reference user corresponding to the ith reference user in the L-bit reference clients, and for Ui N first feature vectors obtained by feature extraction are expressed as { C }1 ,C2 ,……,CN In addition, to Ui The second feature vector obtained by feature extraction of the product information of the purchased financial product is expressed as Vi 。
In an alternative embodiment, fig. 2 is a flowchart of an alternative method of generating a target model from N first feature vectors and second feature vectors according to an embodiment of the present application.
Step S201, clustering operation is carried out on the N first feature vectors, and P vector sets are obtained.
In step S201, each vector set includes the same Q first feature vectors, where P and Q are positive integers.
Step S202, selecting a first feature vector from each vector set as a first target feature vector corresponding to the vector set, and obtaining P first target feature vectors.
Step S203, S first mapping relations are determined according to the P first target feature vectors and the second feature vectors corresponding to each of the L reference users.
In step S203, S is a positive integer, and each first mapping relationship is used to represent a mapping relationship between one first target feature vector of the P first target feature vectors and a second feature vector corresponding to one reference user of the L reference users.
Step S204, the first target feature vector and the second feature vector corresponding to each first mapping relation in the S first mapping relations are spliced to obtain a third feature vector corresponding to the first mapping relation.
Step S205, a target model is generated according to the third feature vector corresponding to each of the S first mapping relations.
Optionally, assume that a corresponding first target feature vector in a certain first mapping relationship is C2 The corresponding second feature vector in the first mapping relation is Vi ，Vi For characterizing the feature information corresponding to the i-th financial product purchased by the reference user, the third feature vector corresponding to the first mapping relationship is C2 |Vi 。
In an alternative embodiment, the determining system inputs a third feature vector corresponding to each of the S first mapping relationships to the generating countermeasure network, where the generating countermeasure network includes a generator and a discriminator, the generator is configured to generate user attribute information corresponding to the virtual user according to the third feature vector, the discriminator is configured to determine similarity between the virtual user and the real user according to the user attribute information generated by the generator, and then perform multiple countermeasure training on the generator and the discriminator according to the similarity between the virtual user and the real user until the similarity between the virtual user and the real user, determined by the target discriminator according to the user attribute information generated by the target generator, is greater than a preset threshold, the target discriminator is a generator after multiple countermeasure training, the target generator is a generator after multiple countermeasure training, and finally, the target model is generated based on the target generator and the target discriminator.
Optionally, the determining system performs clustering operation on the S first mapping relationships according to the second feature vector corresponding to each of the S first mapping relationships to obtain T relationship sets, each relationship set includes N first mapping relationships, T is a positive integer, the second feature vectors corresponding to each of the T first mapping relationships included in each relationship set are the same, the generator includes N sub-generators, the determining system allocates one sub-generator for each of the N first mapping relationships, and inputs the third feature vector corresponding to the first mapping relationship into the corresponding sub-generator in the first mapping relationship to obtain user attribute information corresponding to one virtual user corresponding to the sub-generator, then, the user attribute information corresponding to the N virtual users corresponding to the N sub-generators is input to a discriminator, and the discriminator is used for calculating the probability that the virtual user is a real user, where the calculating method of the probability P is shown in the following formula (1):
optionally, the objective function corresponding to the objective generator is shown in the following formula (2):
JG ＝argmin∑i (log(1-d(G|Vi ))) (2)
optionally, the objective function corresponding to the objective arbiter is as shown in the following formula (3):
JD ＝argmax∑i (log(d(G|Vi ))+log(1-d(G|Vi ))) (3)
In the above formula, d (G|Vi ) And the N user attribute information is used for representing N user attribute information corresponding to the virtual user purchasing the ith financial product, which is generated by the N sub-generators.
In an alternative embodiment, the determining system determines a financial tag corresponding to a target financial product according to product information of the target financial product, wherein the financial tag is used for representing a product type corresponding to the target financial product, then inputs the financial tag to each target sub-generator in the N target sub-generators to obtain user attribute information of a user to be predicted corresponding to each target sub-generator, the target generator comprises N target sub-generators, wherein the target sub-generators are used for predicting one type of user attribute information, and then determines a user portrait to be predicted according to the user attribute information of the user to be predicted corresponding to each target sub-generator in the N target sub-generators.
Optionally, the nth user attribute information corresponding to the user to be predicted generated by the nth target sub-generator is shown in the following formula (4):
gn ＝f(Wn *[C|V]+bn ) (4)
in the formula (4), Wn For characterizing the weight parameters corresponding to the nth target sub-generator, bn For the bias term corresponding to the nth target sub-generator, f is an activation function, and c|v is used to characterize the third feature vector input to the nth target sub-generator.
In an optional embodiment, the N pieces of user attribute information in the determining system at least include region attribute information, behavior attribute information, and risk attribute information, where the region attribute information is used to characterize region information of the user, the behavior attribute information is used to characterize consumption behavior information of the user, and the risk attribute information is used to characterize credit risk information of the user.
Optionally, the region attribute information includes, but is not limited to, consumption location information corresponding to the user and location information of the user, the behavior attribute information includes, but is not limited to, consumption mode information corresponding to the user and consumption credit rating information, and the risk attribute information includes, but is not limited to, credit rating information corresponding to the user and loan information.
In an alternative embodiment, the determining system determines N sub-similarities between the candidate user portrait of each candidate user and the user portrait to be predicted according to N user attribute information corresponding to each candidate user and N user attribute information corresponding to the user to be predicted, where each sub-similarity is used to characterize a similarity between one user attribute information corresponding to the candidate user and one user attribute information corresponding to the user to be predicted, two user attribute information corresponding to each sub-similarity belong to the same type of attribute information, and then determines a similarity between the candidate user portrait and the user portrait to be predicted according to N sub-similarities between the candidate user portrait of each candidate user and the user portrait to be predicted.
Optionally, the determining system firstly obtains a preset weight corresponding to each sub-similarity in the N sub-similarities, and then performs weighted summation on the N sub-similarities according to the preset weight corresponding to each sub-similarity to obtain the candidate user portrait and the user to be predictedSimilarity between images, kth candidate user Uk Corresponding alternative user portrayalWith the user U to be predictedG Corresponding user representation to be predicted->The calculation formula of the similarity between the two is shown in the following formula (5):
in the method, product information of a target financial product is firstly input into a target model to obtain a user portrait to be predicted corresponding to the target financial product, wherein the target model is a neural network model obtained by training according to reference user images corresponding to each of L reference users, each reference user image comprises N user attribute information corresponding to the reference user, the user portrait to be predicted comprises N user attribute information corresponding to the user to be predicted, the reference user is a user who purchases any financial product, the user to be predicted is a user who has purchase intention for the target financial product, L and N are positive integers, then M alternative user portraits corresponding to M alternative users are obtained, M is a positive integer, each alternative user portraits in the M alternative user images comprises N user attribute information corresponding to one alternative user in the M alternative users, each alternative user is a user who does not purchase the target financial product, then the degree between the user portrait to be predicted and each alternative user image in the M alternative user images is detected, and the degree of similarity between the target financial product and at least one target financial product is determined according to the terminal similarity of each alternative user image.
From the above, it can be seen that, the target model in the present application is obtained by training according to the reference user portrait corresponding to each reference user in the L reference users, and the reference user portrait includes N user attribute information corresponding to the reference user, so that it can be deduced that the user portrait to be predicted in the present application is obtained by training the plurality of user attribute information included in the reference user portrait corresponding to each reference user, and compared with the manner of determining the user corresponding to the financial product based on a single user consumption feature in the prior art, the present application does not determine the target user corresponding to the financial product according to a single feature of the candidate user, but determines the target user corresponding to the financial product by judging the similarity between the candidate user portrait and the user portrait to be predicted, thereby achieving the purpose of improving the accuracy when determining the user having the purchase intention for the target financial product.
Therefore, the technical scheme of the application utilizes the target model to generate the user portrait to be predicted, and determines the target user corresponding to the target financial product according to the similarity between the user portrait to be predicted and each candidate user, thereby realizing the purpose of improving the accuracy of determining the user with the purchase intention on the target financial product, further realizing the technical effect of improving the determination efficiency of determining the user with the purchase intention on the target financial product, and further solving the technical problem of low determination efficiency when determining the user with the purchase intention on the financial product in the prior art.
Example 2
According to an embodiment of the present application, an embodiment of a financial product based user determination device is provided. FIG. 3 is a schematic diagram of an alternative financial product based user identification device, as shown in FIG. 3, according to an embodiment of the present application, including: an input unit 301, an acquisition unit 302, and a detection unit 303, and a determination unit 304.
Optionally, the input unit is configured to input product information of a target financial product into the target model to obtain a user portrait to be predicted corresponding to the target financial product, where the target model is a neural network model obtained by training a reference user image corresponding to each reference user of the L reference users, each reference user image includes N user attribute information corresponding to the reference user, the user portrait to be predicted includes N user attribute information corresponding to the user to be predicted, the reference user is a user who purchases any financial product, the user to be predicted is a user who has a purchase intention for the target financial product, L and N are both positive integers, the obtaining unit is configured to obtain M candidate user portraits corresponding to M candidate users, where M is a positive integer, each candidate user portraits in the M candidate user portraits includes N user attribute information corresponding to one candidate user of the M candidate users, the candidate user is a user who does not purchase the target financial product, the detecting unit is configured to detect a degree of similarity between the user portrait to be predicted and each candidate user in the M candidate user portraits, and the target user portraits is determined from the terminal, and the similarity is determined from the target user equipment.
In an alternative embodiment, the financial product-based user determination apparatus further comprises: a first feature extraction unit, a second feature extraction unit and a generation unit.
Optionally, the first feature extraction unit is configured to perform feature extraction on a reference user image corresponding to each reference user in the L reference users to obtain N first feature vectors corresponding to the reference user, where each first feature vector corresponds to user attribute information of one reference user, and the second feature extraction unit is configured to perform feature extraction on product information of a financial product purchased by each reference user to obtain a second feature vector corresponding to the reference user, where the second feature vector corresponding to each reference user is used to characterize a product category corresponding to the financial product purchased by the reference user, and the generating unit is configured to generate the target model according to the N first feature vectors and the second feature vector.
In an alternative embodiment, the generating unit further comprises: the system comprises a first operation subunit, a first determination subunit, a second determination subunit, a splicing subunit and a generation subunit.
Optionally, the first operation subunit is configured to perform clustering operation on N first feature vectors to obtain P vector sets, where each vector set includes the same Q first feature vectors, P and Q are positive integers, the first determination subunit is configured to select one first feature vector from each vector set as a first target feature vector corresponding to the vector set, obtain P first target feature vectors, the second determination subunit is configured to determine S first mapping relations according to the P first target feature vectors and second feature vectors corresponding to each reference user in the L reference users, S is a positive integer, each first mapping relation is used to represent a mapping relation between one first target feature vector in the P first target feature vectors and a second feature vector corresponding to one reference user in the L reference users, and the stitching subunit is configured to stitch the first target feature vector and the second feature vector corresponding to each first mapping relation in the S first mapping relations, and generate a third mapping relation according to the first mapping relation, and generate a third mapping relation corresponding to the first mapping relation.
In an alternative embodiment, the generating subunit further comprises: the system comprises an input module, an countermeasure training module and a generation module.
Optionally, the input module is configured to input a third feature vector corresponding to each of the S first mapping relationships to the generating countermeasure network, where the generating countermeasure network includes a generator and a discriminator, the generator is configured to generate user attribute information corresponding to the virtual user according to the third feature vector, the discriminator is configured to determine similarity between the virtual user and the real user according to the user attribute information generated by the generator, and the countermeasure training module is configured to perform multiple countermeasure training on the generator and the discriminator according to the similarity between the virtual user and the real user until the similarity between the virtual user and the real user, determined by the target discriminator according to the user attribute information generated by the target generator, is greater than a preset threshold, the target discriminator is a generator after multiple countermeasure training, and the target generator is a generator after multiple countermeasure training, and the generating module is configured to generate the target model based on the target generator and the target discriminator.
In an alternative embodiment, the input unit further comprises: a third determination subunit, an input subunit, and a fourth determination subunit.
Optionally, the third determining subunit is configured to determine a financial tag corresponding to the target financial product according to product information of the target financial product, where the financial tag is used to characterize a product type corresponding to the target financial product, the input subunit is configured to input the financial tag to each of N target sub-generators to obtain user attribute information of a user to be predicted corresponding to each target sub-generator, and the target generator includes N target sub-generators, where the target sub-generator is configured to predict one type of user attribute information, and the fourth determining subunit is configured to determine a user portrait to be predicted according to user attribute information of a user to be predicted corresponding to each target sub-generator in the N target sub-generators.
In an optional embodiment, the N pieces of user attribute information in the financial product-based user determining device at least include region attribute information, behavior attribute information, and risk attribute information, where the region attribute information is used to characterize region information of the user, the behavior attribute information is used to characterize consumption behavior information of the user, and the risk attribute information is used to characterize credit risk information of the user.
In an alternative embodiment, the determining unit further comprises: a fifth determination subunit and a sixth determination subunit.
Optionally, the fifth determining subunit is configured to determine N sub-similarities between the candidate user portrait of each candidate user and the user portrait to be predicted according to N user attribute information corresponding to each candidate user and N user attribute information corresponding to the user to be predicted, where each sub-similarity is used to characterize a similarity between one user attribute information corresponding to the candidate user and one user attribute information corresponding to the user to be predicted, and two user attribute information corresponding to each sub-similarity belong to the same class of attribute information, and the sixth determining subunit is configured to determine a similarity between the candidate user portrait and the user portrait to be predicted according to N sub-similarities between the candidate user portrait of each candidate user and the user portrait to be predicted.
In the method, product information of a target financial product is firstly input into a target model to obtain a user portrait to be predicted corresponding to the target financial product, wherein the target model is a neural network model obtained by training according to reference user images corresponding to each of L reference users, each reference user image comprises N user attribute information corresponding to the reference user, the user portrait to be predicted comprises N user attribute information corresponding to the user to be predicted, the reference user is a user who purchases any financial product, the user to be predicted is a user who has purchase intention for the target financial product, L and N are positive integers, then M alternative user portraits corresponding to M alternative users are obtained, M is a positive integer, each alternative user portraits in the M alternative user images comprises N user attribute information corresponding to one alternative user in the M alternative users, each alternative user is a user who does not purchase the target financial product, then the degree between the user portrait to be predicted and each alternative user image in the M alternative user images is detected, and the degree of similarity between the target financial product and at least one target financial product is determined according to the terminal similarity of each alternative user image.
From the above, it can be seen that, the target model in the present application is obtained by training according to the reference user portrait corresponding to each reference user in the L reference users, and the reference user portrait includes N user attribute information corresponding to the reference user, so that it can be deduced that the user portrait to be predicted in the present application is obtained by training the plurality of user attribute information included in the reference user portrait corresponding to each reference user, and compared with the manner of determining the user corresponding to the financial product based on a single user consumption feature in the prior art, the present application does not determine the target user corresponding to the financial product according to a single feature of the candidate user, but determines the target user corresponding to the financial product by judging the similarity between the candidate user portrait and the user portrait to be predicted, thereby achieving the purpose of improving the accuracy when determining the user having the purchase intention for the target financial product.
Therefore, the technical scheme of the application utilizes the target model to generate the user portrait to be predicted, and determines the target user corresponding to the target financial product according to the similarity between the user portrait to be predicted and each candidate user, thereby realizing the purpose of improving the accuracy of determining the user with the purchase intention on the target financial product, further realizing the technical effect of improving the determination efficiency of determining the user with the purchase intention on the target financial product, and further solving the technical problem of low determination efficiency when determining the user with the purchase intention on the financial product in the prior art.
Example 3
According to another aspect of the embodiments of the present application, there is also provided an electronic device, including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the financial product based user determination method of any of the above embodiments 1 via execution of the executable instructions.
Fig. 4 is a schematic diagram of an alternative electronic device according to an embodiment of the present application, and as shown in fig. 4, the embodiment of the present application provides an electronic device, where the electronic device includes a processor, a memory, and a program stored on the memory and executable on the processor, and the processor implements the financial product-based user determination method in any one of the foregoing embodiments 1 when executing the program.
The foregoing embodiment numbers of the present application are merely for describing, and do not represent advantages or disadvantages of the embodiments.
In the foregoing embodiments of the present application, the descriptions of the embodiments are emphasized, and for a portion of this disclosure that is not described in detail in this embodiment, reference is made to the related descriptions of other embodiments.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory. The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash memory (flashRAM). Memory is an example of a computer-readable medium.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transshipment) such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises an element.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.