Financial service recommendation method and device, equipment, medium and program product

Abstract

The invention provides a financial service recommendation method and device, equipment, a medium and a program product, and relates to the field of financial science and technology and the technical field of artificial intelligence. Inputting first state information of the user into a pre-trained first recommendation model to obtain a first recommendation service type output by the first recommendation model, the first state information at least comprising a user type, a service type of a historical recommendation service, and whether the service type of the historical recommendation service is successfully recommended; inputting the first recommendation service type and second state information of the user into a pre-trained second recommendation model to obtain first recommendation service information which is output by the second recommendation model and aims at the first recommendation service type, the second state information at least comprises a user type, a historical recommendation service type, historical recommendation service information and whether the historical recommendation service information is successfully recommended or not; the method improves the accuracy of financial service recommendation.

Description

Technical Field

[0001] The present application relates to the fields of financial technology and artificial intelligence technology, and in particular to a method, device, equipment, medium and program product for recommending financial services. Background Art

[0002] With the continuous development of the economy and the acceleration of digital transformation, enterprises and individuals are increasingly in need of financial products and services that can optimize fund management, provide credit solutions, promote investment growth and risk management. Therefore, recommending financial services suitable for users can not only help users achieve steady growth of wealth and effective risk control, but also promote the continued prosperity and development of the financial industry.

[0003] Currently, the recommendation method for financial services mainly relies on preset rules, for example, screening and recommending corresponding financial products to users based on the popularity of the services.

[0004] However, the method of recommending financial services using preset rules is relatively simple and lacks flexibility, and there is a problem of low accuracy in financial service recommendations. Summary of the invention

[0005] The present application provides a financial service recommendation method, device, equipment, medium and program product, which are used to solve the technical problem of low accuracy of financial service recommendation.

[0006] In a first aspect, the present application provides a method for recommending financial services, comprising:

[0007] Inputting the first state information of the user into a pre-trained first recommendation model to obtain a first recommended service type output by the first recommendation model, wherein the first state information at least includes the user type, the service type of the historically recommended service, and whether the service type of the historically recommended service is successfully recommended;

[0008] The first recommended business type and the user's second status information are input into a pre-trained second recommendation model to obtain the first recommended business information for the first recommended business type output by the second recommendation model, wherein the second status information at least includes the user type, the type of historically recommended business, the historically recommended business information, and whether the historically recommended business information is successfully recommended.

[0009] In a second aspect, the present application provides a training method for a recommendation model for financial services, comprising:

[0010] Acquire a first training sample set, wherein the first training sample includes a first sample state, a first action, and a first result, wherein the first sample state includes at least a user type, a service type of a historically recommended service, and whether the service type of the historically recommended service is successfully recommended, the first action indicates a recommended service type, and the first result indicates success or failure;

[0011] Inputting the first sample state into a first recommendation model to obtain a predicted recommended business type output by the first recommendation model, wherein the first recommendation model is a Markov model;

[0012] Adjusting parameters of the first recommendation model based on the expected cumulative reward corresponding to the predicted recommended business type and the actual cumulative reward corresponding to the first action and the first result;

[0013] Acquire a second training sample set, wherein the second training sample includes a second sample state, a second action, and a second result, wherein the second sample state includes at least a recommended service type, a user type, a type of a historically recommended service, historically recommended service information, and whether the historically recommended service information is successfully recommended, the first action indicates recommended service information, and the second result indicates success or failure;

[0014] Inputting the second sample state into the first recommendation model to obtain predicted recommended business information output by the second recommendation model, where the second recommendation model is a Markov model;

[0015] Based on the expected cumulative reward corresponding to the predicted recommended business information and the actual cumulative reward corresponding to the second action and the second result, the parameters of the second recommendation model are adjusted.

[0016] In a third aspect, the present application provides a financial service recommendation device, including:

[0017] a processing module, configured to input first state information of the user into a pre-trained first recommendation model, and obtain a first recommended service type output by the first recommendation model, wherein the first state information at least includes a user type, a service type of a historically recommended service, and whether the service type of the historically recommended service is successfully recommended;

[0018] The processing module is also used to input the first recommended business type and the user's second status information into a pre-trained second recommendation model to obtain the first recommended business information for the first recommended business type output by the second recommendation model, and the second status information at least includes the user type, the type of historically recommended business, the historically recommended business information, and whether the historically recommended business information is successfully recommended.

[0019] In a fourth aspect, the present application provides a training device for a recommendation model of a financial service, comprising:

[0020] an acquisition module, configured to acquire a first training sample set, wherein the first training sample includes a first sample state, a first action, and a first result, wherein the first sample state includes at least a user type, a service type of a historically recommended service, and whether the service type of the historically recommended service is successfully recommended, the first action indicates a recommended service type, and the first result indicates success or failure;

[0021] a determination module, configured to input the first sample state into a first recommendation model to obtain a predicted recommended business type output by the first recommendation model, wherein the first recommendation model is a Markov model;

[0022] A processing module, configured to adjust parameters of the first recommendation model based on the expected cumulative reward corresponding to the predicted recommended business type and the actual cumulative reward corresponding to the first action and the first result;

[0023] The acquisition module is further used to acquire a second training sample set, wherein the second training sample includes a second sample state, a second action, and a second result, wherein the second sample state includes at least a recommended service type, a user type, a type of a historically recommended service, historically recommended service information, and whether the historically recommended service information is successfully recommended, the first action indicates recommended service information, and the second result indicates success or failure;

[0024] The determination module is further configured to input the second sample state into the first recommendation model to obtain the predicted recommended business information output by the second recommendation model, wherein the second recommendation model is a Markov model;

[0025] The processing module is further used to adjust the parameters of the second recommendation model based on the expected cumulative reward corresponding to the predicted recommended business information and the real cumulative reward corresponding to the second action and the second result.

[0026] In a fifth aspect, an embodiment of the present application provides a financial service recommendation device, including: a memory, a processor;

[0027] The memory stores computer-executable instructions;

[0028] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above-mentioned first aspect and / or various possible implementations of the first aspect and / or the second aspect and / or various possible implementations of the second aspect.

[0029] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-readable storage medium is stored computer execution instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect and / or various possible implementations of the second aspect and / or the second aspect.

[0030] In a seventh aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementations of the first aspect and / or various possible implementations of the second aspect and / or the second aspect.

[0031] The financial service recommendation method, device, equipment, medium and program product provided in the present application input the user's first state information into a pre-trained first recommendation model to obtain the first recommended service type output by the first recommendation model, and the first state information includes at least the user type, the service type of the historically recommended service, and whether the service type of the historically recommended service is successfully recommended. The first recommended service type and the user's second state information are input into a pre-trained second recommendation model to obtain the first recommended service information for the first recommended service type output by the second recommendation model, and the second state information includes at least the user type, the type of the historically recommended service, the historically recommended service information, and whether the historically recommended service information is successfully recommended. The method improves the accuracy of financial service recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0033] Figure 1 A flowchart of a recommended method for financial services provided for this application;

[0034] Figure 2 A flowchart of a training method for a recommendation model for financial services provided in this application;

[0035] Figure 3 A schematic diagram of the structure of the recommendation device for financial services provided by this application;

[0036] Figure 4 A schematic diagram of the structure of a training device for a recommendation model of financial services provided in this application;

[0037] Figure 5 A schematic diagram of the structure of the recommended equipment for the financial services provided in this application.

[0038] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0039] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0040] It should be noted that the financial service recommendation method, device, equipment, medium and program product provided in this application can be used in the fields of financial technology and artificial intelligence technology, and can also be used in any field outside the fields of financial technology and artificial intelligence technology. This application does not limit the application field of the financial service recommendation method, device, equipment, medium and program product.

[0041] With the booming economy and the deepening digital transformation, enterprises and individuals have an increasingly strong demand for financial products and services that can optimize fund management, provide credit solutions, promote investment growth, and effectively manage risks. Therefore, recommending suitable financial services to users can not only help them achieve steady growth of wealth and proper control of risks, but also provide strong support for the continued prosperity and progress of the financial industry.

[0042] Currently, the recommendation method for financial services mainly relies on preset rules, for example, screening and recommending corresponding financial products to users based on the popularity of the services.

[0043] However, the method of recommending financial services using preset rules is relatively simple and lacks flexibility, and there is a problem of low accuracy in financial service recommendations.

[0044] In response to the above problems, the financial service recommendation method provided by the present application constructs a two-level recommendation model. First, the user's first state information (including user type, historical recommended service type and its success or failure) is input into the pre-trained first recommendation model to preliminarily screen out the service type suitable for the user; then, the screened service type and the user's second state information (including user type, historical recommended service type, specific historical recommended service information and its success or failure) are input into the pre-trained second recommendation model to accurately match and output the first recommended service information for the first recommended service type. This method significantly improves the accuracy of financial service recommendations.

[0045] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0046] Figure 1 This is a flow chart of the method for recommending financial services provided by this application. The execution subject of this embodiment is, for example, a financial service recommendation system. Figure 1 As shown, the method includes:

[0047] S101: Input the user's first state information into a pre-trained first recommendation model to obtain a first recommended business type output by the first recommendation model, wherein the first state information at least includes the user type, the business type of the historically recommended business, and whether the business type of the historically recommended business is successfully recommended.

[0048] Among them, user types are classified based on multiple dimensions such as user assets, user credit, user occupation, and financial services that users have handled. They are used to distinguish and locate user groups for financial service recommendations. Different user types have different characteristics and needs, so the recommended financial services will also vary.

[0049] Historically recommended services include financial products (such as credit cards, savings cards), financial management insurance, retail products, etc. Different users of the same user type may receive service recommendations of at least one service type, and the service types recommended to different users may be different.

[0050] The pre-trained first recommendation model is used to determine a first recommended service type according to the first state information of the user.

[0051] It can be understood that, first, the user's first state information (including user type, business type of historically recommended business, and whether the business type of historically recommended business is recommended successfully) is input into the pre-trained first recommendation model, and the pre-trained first recommendation model will predict and analyze the user's needs based on the internal algorithm and training data. Finally, one or more first recommended business types will be output. The first recommended business type is calculated and screened based on the user's current state and information. The recommendation system for financial services can display these recommendations to users and guide users to select the business type of interest, thereby completing the entire recommendation process.

[0052] Optionally, when a user registers for the first time (including status information such as user type), the financial service recommendation system will automatically set the service type with the most recommendations in the historical recommended services corresponding to the user's user type as the first recommended service type. Alternatively, the financial service recommendation system may pre-set a service type as the user's first recommended service type.

[0053] S102: Input the first recommended service type and the second status information of the user into a pre-trained second recommendation model to obtain the first recommended service information for the first recommended service type output by the second recommendation model, and the second status information at least includes the user type, the type of historically recommended services, the historically recommended service information, and whether the historically recommended service information is successfully recommended.

[0054] The pre-trained second recommendation model is used to determine first recommended service information for the first recommended service type according to the first recommended service type and the second state information of the user.

[0055] It can be understood that, first, the first recommended business type and the user's second state information (including user type, type of historically recommended business, historically recommended business information, and whether the historically recommended business information is recommended successfully) are input into the pre-trained first recommendation model, and the pre-trained first recommendation model will predict and analyze the user's needs based on the internal algorithm and training data. Finally, one or more first recommended business information will be output, and the first recommended business information is calculated and screened based on the user's current state and information. The recommendation system for financial services can display these recommendations to users and guide users to select business information of interest, thereby completing the entire recommendation process.

[0056] Optionally, when a user registers for the first time (including status information such as user type), the financial service recommendation system will automatically select the service information with the most recommendations among users who match the user type and have the same first recommended service type as the user's first recommended service information. Alternatively, the financial service recommendation system may preset a piece of service information directly as the user's first recommended service information.

[0057] The financial service recommendation method provided in this embodiment inputs the user's first state information into a pre-trained first recommendation model to obtain a first recommended service type output by the first recommendation model, wherein the first state information includes at least the user type, the service type of historically recommended services, and whether the service type of historically recommended services is successfully recommended; the first recommended service type and the user's second state information are input into a pre-trained second recommendation model to obtain first recommended service information for the first recommended service type output by the second recommendation model, wherein the second state information includes at least the user type, the type of historically recommended services, historically recommended service information, and whether the historically recommended service information is successfully recommended; this method improves the accuracy of financial service recommendations.

[0058] In a possible implementation manner, after the first recommended service information is confirmed, the following steps are included:

[0059] Obtaining a recommendation execution result corresponding to the first recommended business information, and determining whether a recommendation termination condition is met based on the recommendation execution result;

[0060] If the recommendation termination condition is not met, obtaining the updated first state information and the updated second state information of the user;

[0061] Inputting the updated first status information into the first recommendation model to obtain a second recommended business type output by the first recommendation model;

[0062] The second recommended service type and the updated second status information are input into the second recommendation model to obtain second recommended service information for the second recommended service type output by the second recommendation model.

[0063] It is understandable that if the recommendation execution result does not meet the recommendation termination condition, the financial service recommendation system will then obtain the user's latest status information, including the updated first status information and the updated second status information. Then, the financial service recommendation system will input the updated first status information and the updated second status information into different recommendation models to generate new recommended business types and new recommended business information. Specifically, the updated first status information will be input into the first recommendation model to obtain the second recommended business type; then, the second recommended business type and the updated second status information will be input into the second recommendation model to generate the second recommended business information for the second recommended business type.

[0064] In a possible implementation, determining whether a recommendation termination condition is met based on the recommendation execution result includes:

[0065] The recommendation execution result includes budget consumption. If the budget consumption does not reach the budget threshold, it is determined that the recommendation termination condition is not met; or,

[0066] The recommendation execution result includes a recommendation result. If the recommendation result is rejection, it is determined that the recommendation termination condition is not met; or,

[0067] The recommendation execution result includes the number of recommendations. If the number of recommendations does not reach the number threshold, it is determined that the recommendation termination condition is not met.

[0068] Among them, budget consumption refers to the total amount of various expenses incurred in the process of recommending financial services to users, including but not limited to telephone communication fees, gift giving costs, and transportation costs incurred for business recommendations (such as taxi fares).

[0069] The budget threshold is a cost cap set based on different user types (such as VIP customers, ordinary customers, etc.) and the specific financial services that the user plans to handle (such as loans, financial management, insurance, etc.).

[0070] The recommendation result is used to represent the user's feedback on the financial service recommendation, which is specifically divided into two states: acceptance and rejection, to indicate whether the user has adopted the recommended financial service.

[0071] The number of recommendations refers to the total number of times financial services are recommended to users within a preset time period (e.g., the past six months), and is used to measure the frequency of recommendation activities.

[0072] The frequency threshold is a fixed value that can be set to limit the maximum number of times financial services are recommended to users within a specific time period (for example, the past six months). For example, it can be set to no more than five times to ensure that the recommendation activities are not too frequent.

[0073] Understandably, the recommendation execution result comprehensively considers three dimensions: budget consumption, recommendation results and number of recommendations: if the total amount of various expenses incurred in recommending financial services to users (i.e. budget consumption) has not yet reached the cost cap (i.e. budget threshold) set according to user type and business type, or the user's feedback on the recommended business is rejection rather than acceptance, or the number of recommendations within a preset time period (such as the past six months) has not reached the set maximum value (i.e. number threshold), then it can be determined that the termination conditions of the recommendation process are not met.

[0074] In a possible implementation manner, the first recommended service information includes the transaction amount and / or transaction time of the first recommended service.

[0075] Among them, the first recommended business includes but is not limited to: credit cards, savings cards, financial management insurance and gold bars.

[0076] Figure 2 This is a flow chart of the training method of the recommendation model for financial services provided in this application. The execution subject of this embodiment is, for example, a recommendation system for financial services. Figure 2 As shown, the method includes:

[0077] S201: Obtain a first training sample set, where the first training sample includes a first sample state, a first action and a first result. The first sample state includes at least a user type, a business type of a historically recommended business, and whether the business type of a historically recommended business is successfully recommended. The first action indicates a recommended business type, and the first result indicates success or failure.

[0078] S202: Inputting the first sample state into a first recommendation model to obtain a predicted recommended business type output by the first recommendation model, where the first recommendation model is a Markov model.

[0079] S203: Adjusting parameters of the first recommendation model based on the expected cumulative reward corresponding to the predicted recommended business type and the actual cumulative reward corresponding to the first action and the first result.

[0080] Among them, the Markov model is a prediction model based on the Markov property, that is, the current state is only related to its previous state, but not to other states. In financial service recommendation, this means that the user's next recommended service type only depends on the user's current state (such as user type, service type of historical recommended services, and whether the recommendation is successful or not).

[0081] It can be understood that by obtaining a first training sample set, which includes user types, historically recommended business types and first sample states of their success or failure, the first action of the recommended business type, and the first result of success or failure, the first sample state is then input into the Markov model (first recommendation model) to predict the recommended business type, and based on the expected cumulative reward of the predicted recommended business type and the actual cumulative reward of actually executing the first action and the first result, the states, actions and results in the sample set are processed one by one to adjust and optimize the parameters of the first recommendation model.

[0082] The specific implementation formula for adjusting the parameters of the first recommendation model is:

[0083] The calculation formula for expected cumulative rewards is:

[0084]

[0085] The calculation formula for the actual cumulative reward is:

[0086]

[0087] The loss function is calculated as:

[0088]

[0089] By minimizing the loss function , the parameters of the second recommendation model are adjusted using the gradient descent method:

[0090]

[0091] in, is the current first state; is the current first action; is the current reward function; The first state for the next time; For the first action of the next time; are the parameters of the current first recommended model; is the discount factor, which indicates the importance of future rewards and is usually set between between; is the current reward function, is the learning rate.

[0092] S204: Obtain a second training sample set, where the second training sample includes a second sample state, a second action, and a second result. The second sample state includes at least the recommended business type, the user type, the type of historically recommended business, the historically recommended business information, and whether the historically recommended business information is successfully recommended. The first action indicates the recommended business information, and the second result indicates success or failure.

[0093] S205: Inputting the second sample state into the first recommendation model to obtain the predicted recommended business information output by the second recommendation model, where the second recommendation model is a Markov model.

[0094] S206: Adjust the parameters of the second recommendation model based on the expected cumulative reward corresponding to the predicted recommended business information and the actual cumulative reward corresponding to the second action and the second result.

[0095] It can be understood that by obtaining a second training sample set, which includes the recommended business type, user type, type of historically recommended business, historically recommended business information, a second sample state of whether the historically recommended business information is successfully recommended, a second action of the recommended business type, and a second result of success or failure, the second sample state is then input into the Markov model (second recommendation model) to predict the recommended business information, and based on the expected cumulative reward of predicting the recommended business information and the actual cumulative reward of actually executing the second action and the second result, the states, actions and results in the sample set are processed one by one to adjust and optimize the parameters of the second recommendation model.

[0096] The specific implementation formula for adjusting the parameters of the second recommendation model is similar to the specific implementation formula for adjusting the parameters of the first recommendation model, and will not be elaborated here.

[0097] In one possible implementation, the reward function of the first recommendation model and the reward function of the second recommendation model include user feedback rewards and / or budget consumption rewards, wherein the user feedback rewards are determined based on the success or failure of the recommendation, and the budget consumption rewards are determined based on the budget consumption in the recommendation execution results.

[0098] Among them, user feedback based on successful recommendation is positive feedback (for example, the user accepts the recommended business information), and user feedback based on successful recommendation is negative feedback (for example, the user rejects the recommended business information).

[0099] The calculation formula for user feedback rewards is:

[0100]

[0101] in, For the The state of the second; For the The action of the second is the measurement function of user feedback. (For example, if user feedback is positive feedback, then Can be taken ; If the user feedback is negative, then Can be taken . )

[0102] The calculation formula for budget consumption rewards is:

[0103]

[0104] in, For the current state; for the current action; The current budget consumption of the user.

[0105] In a possible implementation, the reward function of the first recommendation model and the reward function of the second recommendation model are weighted sums of user feedback rewards and budget consumption rewards.

[0106] Among them, the reward function is:

[0107]

[0108] in, The weight of the user feedback reward, The weight of the reward for consuming the budget.

[0109] Figure 3 A schematic diagram of the structure of the recommendation device for the financial service provided by this application, such as Figure 3 As shown, the financial service recommendation device 300 provided in this embodiment includes:

[0110] Processing module 301, used for inputting the first state information of the user into the pre-trained first recommendation model to obtain the first recommended service type output by the first recommendation model, wherein the first state information at least includes the user type, the service type of the historically recommended service, and whether the service type of the historically recommended service is recommended successfully;

[0111] The processing module 301 is also used to input the first recommended business type and the user's second state information into a pre-trained second recommendation model to obtain the first recommended business information for the first recommended business type output by the second recommendation model, and the second state information at least includes the user type, the type of historically recommended business, the historically recommended business information, and whether the historically recommended business information is successfully recommended.

[0112] In a possible implementation, the device further includes: an acquisition module 302 and a determination module 303;

[0113] An acquisition module 302 is used to acquire a recommendation execution result corresponding to the first recommended business information;

[0114] A determination module 303 is used to determine whether a recommendation termination condition is met based on the recommendation execution result;

[0115] The acquisition module 302 is further configured to acquire the updated first state information and the updated second state information of the user if the recommendation termination condition is not met;

[0116] The processing module 301 is further configured to input the updated first status information into the first recommendation model to obtain a second recommended service type output by the first recommendation model;

[0117] The processing module 301 is further configured to input the second recommended service type and the updated second status information into the second recommendation model to obtain second recommended service information output by the second recommendation model for the second recommended service type.

[0118] In a possible implementation, the determination module 303 is further configured to recommend that the execution result includes budget consumption, and if the budget consumption does not reach the budget threshold, determine that the recommendation termination condition is not met; or,

[0119] The determination module 303 is further configured to recommend an execution result including a recommendation result, and if the recommendation result is a rejection, determine that the recommendation termination condition is not met; or,

[0120] The determination module 303 is further configured to recommend that the execution result includes the number of recommendations. If the number of recommendations does not reach a threshold, it is determined that the recommendation termination condition is not met.

[0121] In a possible implementation manner, the determination module 303 is further configured to: the first recommended service information includes a transaction amount and / or a transaction time of the first recommended service.

[0122] The financial service recommendation device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and this embodiment will not be described in detail here.

[0123] Figure 4 A structural diagram of a training device for a recommendation model of financial services provided in this application, such as Figure 4 As shown, the training device 400 for the recommendation model of financial services provided in this embodiment includes:

[0124] An acquisition module 401 is used to acquire a first training sample set, where the first training sample includes a first sample state, a first action, and a first result. The first sample state includes at least a user type, a service type of a historically recommended service, and whether the service type of the historically recommended service is successfully recommended. The first action indicates a recommended service type, and the first result indicates success or failure.

[0125] A determination module 402 is used to input the first sample state into a first recommendation model to obtain a predicted recommended business type output by the first recommendation model, where the first recommendation model is a Markov model;

[0126] The processing module 403 is used to adjust the parameters of the first recommendation model based on the expected cumulative reward corresponding to the predicted recommended business type and the actual cumulative reward corresponding to the first action and the first result;

[0127] The acquisition module 401 is further used to acquire a second training sample set, where the second training sample includes a second sample state, a second action, and a second result, where the second sample state includes at least a recommended service type, a user type, a type of historically recommended services, historically recommended service information, and whether the historically recommended service information is successfully recommended, the first action indicates recommended service information, and the second result indicates success or failure;

[0128] The determination module 402 is further used to input the second sample state into the first recommendation model to obtain the predicted recommended business information output by the second recommendation model, where the second recommendation model is a Markov model;

[0129] The processing module 403 is further configured to adjust the parameters of the second recommendation model based on the expected cumulative reward corresponding to the predicted recommended business information and the actual cumulative reward corresponding to the second action and the second result.

[0130] In one possible implementation, the determination module 402 also includes a reward function for the first recommendation model and a reward function for the second recommendation model, including a user feedback reward and / or a budget consumption reward, wherein the user feedback reward is determined based on the success or failure of the recommendation, and the budget consumption reward is determined based on the budget consumption in the recommendation execution result.

[0131] In a possible implementation, the determination module 402 is further used for the reward function of the first recommendation model and the reward function of the second recommendation model to be a weighted sum of the user feedback reward and the budget consumption reward.

[0132] The training device for the recommendation model of financial services provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar, and are not described in detail in this embodiment.

[0133] Figure 5 This is a schematic diagram of the structure of the recommended equipment for the financial services provided by this application. Figure 5 As shown, the electronic device 500 provided in this embodiment includes: at least one processor 501 and a memory 502. Optionally, the device 500 also includes a communication component 503. The processor 501, the memory 502 and the communication component 503 are connected via a bus 504.

[0134] In a specific implementation process, at least one processor 501 executes the computer-executable instructions stored in the memory 502, so that at least one processor 501 executes the above method.

[0135] The specific implementation process of the processor 501 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.

[0136] In the above embodiments, it should be understood that the processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the invention can be directly implemented as a hardware processor, or can be implemented by a combination of hardware and software modules in the processor.

[0137] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (NVM), such as at least one disk storage.

[0138] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of this application is not limited to only one bus or one type of bus.

[0139] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0140] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0141] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0142] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0143] The division of units is only a logical function division, and there may be other divisions in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0144] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0145] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0146] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0147] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0148] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0149] It should be further noted that, although the various steps in the flowchart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0150] It should be understood that the above-mentioned device embodiments are only illustrative, and the device of the present application can also be implemented in other ways. For example, the division of units / modules in the above-mentioned embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0151] In addition, unless otherwise specified, each functional unit / module in each embodiment of the present application may be integrated into one unit / module, each unit / module may exist physically separately, or two or more units / modules may be integrated together. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of a software program module.

[0152] If the integrated unit / module is implemented in the form of hardware, the hardware may be a digital circuit, an analog circuit, etc. The physical implementation of the hardware structure includes but is not limited to transistors, memristors, etc. Unless otherwise specified, the processor may be any appropriate hardware processor, such as a CPU, a GPU, an FPGA, a DSP, an ASIC, etc. Unless otherwise specified, the storage unit may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory (RRAM), a dynamic random access memory (DRAM), a static random access memory (SRAM), an enhanced dynamic random access memory (EDRAM), a high-bandwidth memory (HBM), a hybrid memory cube (HMC), etc.

[0153] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0154] In the above embodiments, the description of each embodiment has its own emphasis. For the part not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0155] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0156] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for recommending financial services, characterized in that: include: Inputting the first state information of the user into a pre-trained first recommendation model to obtain a first recommended service type output by the first recommendation model, wherein the first state information at least includes the user type, the service type of the historically recommended service, and whether the service type of the historically recommended service is successfully recommended; The first recommended business type and the user's second status information are input into a pre-trained second recommendation model to obtain the first recommended business information for the first recommended business type output by the second recommendation model, wherein the second status information at least includes the user type, the type of historically recommended business, the historically recommended business information, and whether the historically recommended business information is successfully recommended.

2. The method according to claim 1, characterized in that Also includes: Acquire a recommendation execution result corresponding to the first recommended service information, and determine whether a recommendation termination condition is met based on the recommendation execution result; If the recommendation termination condition is not met, obtaining the updated first state information and the updated second state information of the user; Inputting the updated first status information into the first recommendation model to obtain a second recommended business type output by the first recommendation model; The second recommended service type and the updated second status information are input into the second recommendation model to obtain second recommended service information output by the second recommendation model for the second recommended service type.

3. The method according to claim 2, characterized in that The determining whether a recommendation termination condition is met based on the recommendation execution result includes: The recommendation execution result includes budget consumption, and if the budget consumption does not reach the budget threshold, it is determined that the recommendation termination condition is not met; or, The recommendation execution result includes a recommendation result, and if the recommendation result is rejection, it is determined that the recommendation termination condition is not met; or, The recommendation execution result includes the number of recommendations. If the number of recommendations does not reach a threshold, it is determined that the recommendation termination condition is not met.

4. The method according to any one of claims 1 to 3, characterized in that: The first recommended service information includes the transaction amount and / or transaction time of the first recommended service.

5. A training method for a recommendation model for financial services, characterized in that: include: Acquire a first training sample set, wherein the first training sample includes a first sample state, a first action, and a first result, wherein the first sample state includes at least a user type, a service type of a historically recommended service, and whether the service type of the historically recommended service is successfully recommended, the first action indicates a recommended service type, and the first result indicates success or failure; Inputting the first sample state into a first recommendation model to obtain a predicted recommended business type output by the first recommendation model, wherein the first recommendation model is a Markov model; Adjusting parameters of the first recommendation model based on the expected cumulative reward corresponding to the predicted recommended business type and the actual cumulative reward corresponding to the first action and the first result; Acquire a second training sample set, wherein the second training sample includes a second sample state, a second action, and a second result, wherein the second sample state includes at least a recommended service type, a user type, a type of a historically recommended service, historically recommended service information, and whether the historically recommended service information is successfully recommended, the first action indicates recommended service information, and the second result indicates success or failure; Inputting the second sample state into the first recommendation model to obtain predicted recommended business information output by the second recommendation model, where the second recommendation model is a Markov model; Based on the expected cumulative reward corresponding to the predicted recommended business information and the actual cumulative reward corresponding to the second action and the second result, the parameters of the second recommendation model are adjusted.

6. The method according to claim 5, characterized in that The reward function of the first recommendation model and the reward function of the second recommendation model include a user feedback reward and / or a budget consumption reward, wherein the user feedback reward is determined based on the success or failure of the recommendation, and the budget consumption reward is determined based on the budget consumption in the recommendation execution result.

7. The method according to claim 6, characterized in that The reward function of the first recommendation model and the reward function of the second recommendation model are a weighted sum of the user feedback reward and the budget consumption reward.

8. A financial service recommendation device, characterized in that: include: a processing module, configured to input first state information of the user into a pre-trained first recommendation model, and obtain a first recommended service type output by the first recommendation model, wherein the first state information at least includes a user type, a service type of a historically recommended service, and whether the service type of the historically recommended service is successfully recommended; The processing module is also used to input the first recommended business type and the user's second status information into a pre-trained second recommendation model to obtain the first recommended business information for the first recommended business type output by the second recommendation model, and the second status information at least includes the user type, the type of historically recommended business, the historically recommended business information, and whether the historically recommended business information is successfully recommended.

9. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

11. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 7 when being executed by a processor.

Images