Method and device for recommending rights and interests

Abstract

The invention discloses a method and a device for recommending rights and interests, and relates to the technical field of computers. A specific embodiment of the method comprises: in response to a received right and interest recommendation request, determining a target user group and user rights and interests corresponding to each user in the target user group, the user rights and interests being determined from the rights and interests by a pre-constructed decision tree model, a training sample of the decision-making tree model comes from an earning recommendation result of the rights and interests; and recommending the corresponding user rights and interests to each user in the target user group, and obtaining a recommendation result of the rights and interests. According to the embodiment, the efficient and low-cost right and interest recommendation method is achieved, compared with an existing recommendation method, the recommendation cost is reduced, the recommendation reasonability is guaranteed based on the user right and interest recommendation method corresponding to each user obtained on the basis of the decision tree model, the right and interest cancel-after-verification rate is increased, and follow-up right and interest demand prediction is facilitated.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and more particularly to a preferred method and apparatus. Background Technology

[0002] In modern operational systems, offering coupons and discount vouchers to users attracts new users and maintains user engagement. For newly developed benefits, since there's no historical recommendation history to reference, it's impossible to know user demand. Therefore, a random recommendation experiment is used, initially showing newly developed benefits to a subset of users, and then using the redemption data to predict demand for those benefits.

[0003] In the process of realizing this invention, the inventors discovered the following problems in the prior art:

[0004] When multiple new benefits are developed, it is necessary to conduct random experiments on each benefit to predict the demand for each benefit. This will consume a lot of experimental costs. Moreover, since the benefits are randomly recommended, the redemption rate of the benefits is low. Benefits are often recommended to unsuitable users, which is not conducive to subsequent benefit demand prediction. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a method and apparatus for recommending benefits, realizing an efficient and low-cost benefit recommendation method. Compared with existing recommendation methods, it reduces recommendation costs, and the user benefits corresponding to each user obtained based on the decision tree model ensure the rationality of the recommendation, improve the redemption rate of benefits, and help predict subsequent benefit demand.

[0006] To achieve the aforementioned objective, according to one aspect of the present invention, a method for recommending rights is provided, comprising:

[0007] In response to receiving a rights recommendation request, a target user group is determined, as well as the user rights corresponding to each user in the target user group. The user rights are determined from the rights by a pre-built decision tree model, and the training samples of the decision tree model come from the preliminary recommendation results of the rights.

[0008] Recommend the corresponding user benefits to each user in the target user group, and obtain the recommendation results of the benefits.

[0009] Optionally, before determining the user rights corresponding to each user in the target user group, the method further includes: constructing a training sample set based on the preliminary recommendation results of the rights, training a preset decision tree initial model, and obtaining a decision tree model, wherein the decision tree model has the ability to predict the user rights corresponding to all users.

[0010] Optionally, before constructing a training sample set based on the preliminary recommendation results of the benefits, the method further includes: determining a preliminary user group according to preset user selection rules; determining the preliminary user benefits corresponding to each user in the preliminary user group according to the benefits, recommending the preliminary user benefits to the corresponding users in the preliminary user group, and using the obtained benefit redemption data as the preliminary recommendation results of the benefits.

[0011] Optionally, the preliminary user group and the target user group are determined by using a reinforcement learning strategy.

[0012] Optionally, the initial decision tree model is constructed using a random forest model or a gradient stochastic model. The constituent features of the initial decision tree model include: the user's basic features, the historical order features of the trading platform, and the historical browsing features of the trading platform.

[0013] Optionally, after obtaining the recommendation result of the right, the method further includes: if the recommendation result of the right meets a preset result threshold, persistently storing the recommendation result of the right.

[0014] Optionally, if the recommendation result of the benefit does not meet the result threshold, the method further includes: repeatedly executing the following steps until the recommendation result of the benefit meets the result threshold: constructing a new training sample set based on the recommendation result of the benefit, training the decision tree model, and obtaining a decision tree optimization model; determining new user benefits corresponding to each user in the target user group based on the decision tree optimization model, recommending the corresponding new user benefits to each user in the target user group, and obtaining a new recommendation result for the benefit.

[0015] According to a second aspect of the present invention, an apparatus for recommending rights is provided, comprising:

[0016] The user rights determination module is used to respond to a received rights recommendation request, determine the target user group, and the user rights corresponding to each user in the target user group. The user rights are determined from the rights by a pre-built decision tree model, and the training samples of the decision tree model come from the preliminary recommendation results of the rights.

[0017] The recommendation result determination module is used to recommend the corresponding user benefits to each user in the target user group and obtain the recommendation results of the benefits.

[0018] According to a third aspect of the present invention, an electronic recommendation device is provided, comprising:

[0019] One or more processors;

[0020] Storage device for storing one or more programs.

[0021] When the one or more programs are executed by the one or more processors, the one or more processors implement the method provided in the first aspect of the embodiments of the present invention.

[0022] According to a fourth aspect of the present invention, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method provided in the first aspect of the present invention.

[0023] One embodiment of the invention has the following advantages or beneficial effects: by responding to a received benefit recommendation request, determining the target user group and the user benefits corresponding to each user in the target user group, the user benefits are determined from the benefits by a pre-constructed decision tree model, and the training samples of the decision tree model come from the preliminary benefit recommendation results; the technical solution of recommending the corresponding user benefits to each user in the target user group and obtaining the benefit recommendation results realizes an efficient and low-cost benefit recommendation method. Compared with existing recommendation methods, it reduces recommendation costs, and the user benefits corresponding to each user obtained based on the decision tree model ensure the rationality of the recommendation, improve the efficiency of the recommendation, increase the benefit redemption rate, and help predict subsequent benefit demand. Attached Figure Description

[0024] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein:

[0025] Figure 1 This is a schematic diagram of the main flow of the recommendation method according to the embodiments of the present invention;

[0026] Figure 2 This is a schematic diagram of a decision tree model according to an embodiment of the present invention;

[0027] Figure 3 This is a detailed flowchart illustrating the recommended method for the benefits of this embodiment;

[0028] Figure 4 This is a schematic diagram of the overall system architecture of the recommended method according to the claims of this invention.

[0029] Figure 5 This is a schematic diagram of the main modules of the recommended device according to an embodiment of the present invention;

[0030] Figure 6 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied;

[0031] Figure 7 This is a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers of the present invention. Detailed Implementation

[0032] It should be noted that the collection, updating, analysis, use, transmission, and storage of user personal information involved in the technical solution of this invention all comply with relevant laws and regulations, are used for legitimate and reasonable purposes, and are not shared, disclosed, or sold outside of these legitimate uses, and are subject to supervision and management by national regulatory authorities. Necessary measures should be taken to selectively block the use or access to personal information data to prevent unauthorized access to such personal information data, ensure that personnel authorized to access personal information data comply with relevant laws and regulations, and ensure the security of user personal information. Furthermore, once this user personal information data is no longer needed, the risk should be minimized by restricting or even prohibiting data collection and / or deleting the data.

[0033] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0034] Existing benefit recommendation methods, when multiple new benefits are developed, require random experiments for each benefit to predict demand for each benefit. This incurs significant experimental costs, and because benefits are randomly recommended, the redemption rate is low, often resulting in benefits being recommended to unsuitable users. This is detrimental to subsequent benefit demand prediction and cannot adequately meet the needs of practical applications.

[0035] To address the aforementioned problems in existing technologies, this invention proposes a rights recommendation method. Based on a decision tree model trained from the preliminary recommendation results of rights, the method obtains the user rights corresponding to each user in the target user group, recommends the corresponding user rights to each user in the target user group, and obtains the rights recommendation results. This achieves efficient and low-cost rights recommendation, reducing recommendation costs compared to existing methods. Moreover, the user rights corresponding to each user obtained based on the decision tree model ensure the rationality of the recommendations, improve recommendation efficiency, and increase the rights redemption rate, which is helpful for subsequent rights demand prediction.

[0036] In the description of the embodiments of the present invention, the terms involved and their meanings are as follows:

[0037] ε-greedy strategy: This is a basic reinforcement learning strategy, mainly used to make trade-offs between exploration and exploitation. In ε-greedy, the agent selects a random action with a probability of ε at each decision point and selects the action that is currently considered the best with a probability of 1-ε.

[0038] Figure 1 This is a schematic diagram of the main flow of the recommendation method according to an embodiment of the present invention, as shown below. Figure 1 As shown, the recommended method according to the embodiments of the present invention includes the following steps S101 to S102.

[0039] Step S101: In response to receiving a rights recommendation request, determine the target user group and the user rights corresponding to each user in the target user group. The user rights are determined from the rights by a pre-constructed decision tree model. The training samples of the decision tree model come from the preliminary recommendation results of the rights.

[0040] Specifically, in this embodiment of the invention, the rights and interests are a concept of a set of rights and interests, representing at least one newly developed rights and interests. The degree of user preference for these rights and interests is unknown to operators or business personnel. In this case, recommending rights and interests to users can be referred to as a cold start for rights and interests. Upon receiving a rights and interests recommendation request, target users are selected from the entire user base according to preset rules for determining the target user group, forming a target user group. Specifically, target users can be randomly selected according to a preset ratio, or they can be selected based on existing probability algorithms, or a specified number of target users can be selected based on user profiles in a data warehouse. This embodiment of the invention does not impose specific limitations. Then, based on the prediction results of user preference rights and interests using a decision tree model, the user rights and interests corresponding to each user in the target user group are predicted, thus obtaining the target user group corresponding to this rights and interests recommendation request, and the user rights and interests corresponding to each user in the target user group.

[0041] According to an embodiment of the present invention, before determining the user rights corresponding to each user in the target user group, the method further includes: constructing a training sample set based on the preliminary recommendation results of the rights, training a preset decision tree initial model, and obtaining a decision tree model, wherein the decision tree model has the ability to predict the user rights corresponding to all users.

[0042] Specifically, this embodiment of the invention utilizes a supervised learning algorithm's decision tree model to predict user rights on the platform. The construction of the decision tree model relies heavily on training sample data for rights. Before determining the user rights corresponding to each user in the target user group, this embodiment first conducts a preliminary recommendation of newly developed rights, such as random rights recommendations for random users, random rights recommendations for designated users, or random rights recommendations for designated users, etc., to obtain preliminary recommendation results. These preliminary recommendation results are then labeled and integrated to obtain a training sample set. This training sample set is used to train a pre-set initial decision tree model, resulting in a decision tree model. The trained decision tree model can be used to predict the user rights of any user on the trading platform, possessing the ability to predict the user rights corresponding to all users.

[0043] According to another embodiment of the present invention, the initial decision tree model is constructed by a random forest model or a gradient stochastic model, and the constituent features of the initial decision tree model include: the user's basic features, the historical order features of the trading platform, and the historical browsing features of the trading platform.

[0044] Specifically, the initial decision tree model in this embodiment of the invention is mainly constructed using a random forest model or a gradient stochastic model, although other commonly used machine learning models can also be used. The features represented by each node of the initial decision tree model mainly include: basic user features, such as account registration city, gender, and age; historical order features of the trading platform, such as the city with the most shipments, the number of orders placed in a recent period, and the order interval; and historical browsing features of the trading platform, such as the number of browsing visits and browsing duration in a recent period. Based on these feature nodes, the initial decision tree model is trained using training samples to obtain the decision tree model.

[0045] Figure 2 This is a schematic diagram of the decision tree model in an embodiment of the present invention. The decision tree model is primarily based on fundamental data from a data warehouse. It constructs and combines nodes with various features at the user level, and then trains using the preliminary recommendation results of benefits to obtain a decision tree model capable of predicting the user benefits for all users. The decision tree model in the diagram starts by searching and matching at each level based on whether the user is a member, and then matches based on the user's delivery city, account registration city, gender, and age characteristics to obtain the matched user benefits.

[0046] According to another embodiment of the present invention, before constructing a training sample set based on the preliminary recommendation results of the rights and benefits, the method further includes: determining a preliminary user group according to a preset user selection rule; determining the preliminary user rights and benefits corresponding to each user in the preliminary user group according to the rights and benefits, recommending the preliminary user rights and benefits to the corresponding users in the preliminary user group, and using the obtained rights and benefits redemption data as the preliminary recommendation results of the rights and benefits.

[0047] Specifically, before constructing the training sample set, a preliminary user group is first determined from the full user base according to preset user selection rules. This user selection rule can be either to randomly select users meeting a preset proportion as the preliminary user group, or to conduct a secondary, finer-grained selection from a randomly selected candidate user group to obtain the preliminary user group. The specific selection rules are configured according to different business scenarios. Then, based on these newly developed benefits, the corresponding user benefits for each user in the preliminary user group are randomly determined, and these determined user benefits are recommended to the corresponding users in the preliminary user group. After a period of accumulation, user redemption data for these newly developed benefits is collected, and this user redemption data is used as the preliminary benefit recommendation result.

[0048] According to another embodiment of the present invention, the preliminary user group and the target user group are determined by using a reinforcement learning strategy.

[0049] Specifically, in this embodiment of the invention, the determination of the preliminary user group and the target user group uses a reinforcement learning ε-greedy strategy. When determining the preliminary user group, a small ε value is selected from (0, 1), for example, ε is 0.1. First, from the first candidate user group randomly determined from all users, for each user in the first candidate user group, a decision is made on whether to recommend benefits to the user with a probability of ε, that is, to determine whether the user is a user in the preliminary user group. For example, a random number is generated, and if the number is less than ε, then the user is a user in the preliminary user group. It can be understood that when ε is small, the number of users included in the preliminary user group is not large. On the one hand, the preliminary recommendation itself is a relatively small-scale preliminary survey, and on the other hand, the cost of benefit recommendation is also considered. After all, benefits such as coupons, discount coupons, and exchange purchases require costs. Too many aimless recommendations not only waste benefit resources but also consume benefit costs. Therefore, using a small ε value is more suitable for the preliminary stage.

[0050] Accordingly, regarding the determination of the target user group, since the user benefits recommended to each user in the target user group at this stage are predicted based on the decision tree model, the recommendation behavior of recommending user benefits to users in the corresponding target user group is a reasoned and purposeful recommendation. Therefore, the target user group is decided with a probability of 1-ε. In the second candidate user group randomly determined from the full user base, for each user in the second candidate user group, a decision is made with a probability of 1-ε to recommend benefits to users in the second candidate user group, that is, to determine whether they are users in the target user group. Taking ε as 0.1 as an example, the decision to recommend benefits to users is made with a probability of 0.9. A random number is generated, and if the number is less than 1-ε, then the user is a user in the target user group. It can be understood that when ε is small, the target user group contains a large number of users, which also meets the needs of benefit recommendation, distributing benefits to suitable users as much as possible.

[0051] By using the ε-greedy strategy of reinforcement learning, the probability of ε is used to explore the preliminary user group, and the probability of 1-ε is used to determine the target user group. Under the premise that the sum of the probabilities of the preliminary user group and the target user group is 1, the exploration rate ε can be flexibly adjusted to adjust the preliminary user group and the target user group to adapt to various scenarios.

[0052] Step S102: Recommend the corresponding user benefits to each user in the target user group, and obtain the recommendation results of the benefits.

[0053] Specifically, based on the target user group determined above and the user rights corresponding to each user in the target user group, the corresponding user rights are recommended to each user in the target user group, and the redemption result of the received user rights by each user in the target user group is obtained, which is the recommendation result of the rights.

[0054] The method first recommends corresponding user benefits to users in the initial user group to obtain the benefit recommendation results. Then, based on the initial user group recommendation results, a training sample is constructed to obtain a decision tree model, which in turn obtains the corresponding user benefits recommended to each user in the target user group, and then recommends the corresponding user benefits to each user in the target user group. This achieves an efficient and low-cost benefit recommendation method, which can effectively ensure the rationality of recommending benefits to the target user group based on only one initial user group recommendation, and improves the benefit redemption rate.

[0055] According to one embodiment of the present invention, after obtaining the recommendation result of the right, the method further includes: if the recommendation result of the right meets a preset result threshold, persistently storing the recommendation result of the right.

[0056] Specifically, to ensure the data quality of the rights recommendation results, after obtaining the recommendation results, it is necessary to verify them to determine whether they meet a preset result threshold, such as the rights redemption rate threshold. If the rights recommendation results meet the preset result threshold, the data quality of the recommendation results meets the requirements, and the recommendation results are persistently stored for subsequent rights demand prediction.

[0057] According to another embodiment of the present invention, if the recommendation result of the benefit does not meet the result threshold, the method further includes: repeatedly performing the following steps until the recommendation result of the benefit meets the result threshold: constructing a new training sample set based on the recommendation result of the benefit, training the decision tree model to obtain a decision tree optimization model; determining new user benefits corresponding to each user in the target user group based on the decision tree optimization model, recommending the corresponding new user benefits to each user in the target user group, and obtaining a new recommendation result for the benefit.

[0058] Specifically, if the recommended benefits do not meet the predetermined threshold, it indicates that the decision tree model needs further training. In this case, using the obtained benefit recommendation results, a new training sample set is constructed to train the decision tree model, resulting in an optimized decision tree model. This optimized model is then used to predict new user benefits for each user in the target user group, and subsequently, the corresponding new user benefits are recommended to each user in the target user group, yielding new recommendation results. It is then determined whether the new recommendation results meet the preset threshold. If they do, the new recommendation results are stored; otherwise, the above steps are repeated to perform a new round of model training, predicting new user benefits, and recommending them to the target user group, until the benefit recommendation results meet the threshold.

[0059] By verifying the recommendation results as described above, not only is the reliability and availability of the data guaranteed, but high-quality basic data is also provided for subsequent prediction of rights and interests.

[0060] Figure 3This is a detailed flowchart illustrating the benefit recommendation method of this embodiment. First, a first candidate user group and a second candidate user group are determined. Using a reinforcement learning ε-greedy strategy, based on the first candidate user group, a random decision is made with ε probability to determine whether to recommend benefits to users, resulting in a preliminary user group. From a set of benefits including benefits A, B, and C, preliminary benefits corresponding to users in this preliminary user group are randomly determined, and these preliminary benefits are recommended to the corresponding users in the preliminary user group, obtaining the preliminary benefit recommendation result. An initial decision tree model is trained based on this preliminary recommendation result to obtain a decision tree model. Based on the second candidate user group, a random decision is made with 1-ε probability to determine whether to recommend benefits to users, resulting in a target user group. Then, based on the decision tree model, the user benefits corresponding to users in the target user group are predicted, and the corresponding user benefits are recommended to users in the target user group, obtaining the benefit recommendation result. Finally, it is determined whether the benefit recommendation result meets a preset result threshold. If the criteria are met, the recommended benefits are persistently saved. If not, the previously determined target user group can be used directly, or a second alternative user group can be used to randomly decide whether to recommend benefits to users again with a probability of 1-ε, resulting in a newly selected target user group. A decision tree model is then trained based on this recommendation result to obtain an optimized decision tree model. This optimized model is used to predict the new user benefits corresponding to users in the current target user group, enabling a new round of recommendations based on these new benefits. At this point, it is determined whether the new benefit recommendation result meets the result threshold. If it does, the new recommendation result is saved and persistently saved. If not, the decision tree model is continuously optimized and trained for a new round of recommendations until the benefit recommendation result meets the result threshold.

[0061] Figure 4 This is a schematic diagram of the overall system architecture of the rights recommendation method according to an embodiment of the present invention. The rights recommendation method of this embodiment can also be visualized as a cold start strategy for rights. A user pool is selected from a user pool through a user selection system to identify a first candidate user group and a second candidate user group. A data warehouse provides rich and complete basic data for the construction of the initial decision tree model and the determination of constituent features in the cold start strategy. Then, based on each newly developed right (right A, right B, right C, ...) in the rights set, the user rights corresponding to users in the target user group are determined, and the corresponding user rights are recommended to each user in the target user group through a rights push system. In this embodiment, the relevant information on recommending the corresponding user rights to each user in the target user group is called the rights recommendation details. The rights push system can also perform secondary screening and control on the received rights recommendation details from aspects such as budget cost. Finally, the rights recommendation details that meet the budget cost requirements are sent to the front-end page to reach each user in the target user group.

[0062] Figure 5 This is a schematic diagram of the main modules of the recommended device according to an embodiment of the present invention. Figure 5 As shown, the rights recommendation device 500 mainly includes a user rights determination module 501 and a recommendation result determination module 502.

[0063] User rights determination module 501 is used to respond to receiving a rights recommendation request, determine a target user group, and the user rights corresponding to each user in the target user group. The user rights are determined from the rights by a pre-built decision tree model. The training samples of the decision tree model come from the preliminary recommendation results of the rights.

[0064] The recommendation result determination module 502 is used to recommend the corresponding user rights to each user in the target user group and obtain the recommendation result of the rights.

[0065] According to an embodiment of the present invention, the rights recommendation device 500 further includes a prediction module (not shown in the figure), which is used to: before determining the user rights corresponding to each user in the target user group, construct a training sample set based on the preliminary recommendation results of the rights, train a preset decision tree initial model, and obtain a decision tree model, wherein the decision tree model has the ability to predict the user rights corresponding to all users.

[0066] According to another embodiment of the present invention, the rights recommendation device 500 further includes a preliminary recommendation result acquisition module (not shown in the figure), which is used to: determine a preliminary user group according to a preset user selection rule before constructing a training sample set based on the preliminary recommendation result of the rights; determine the preliminary user rights corresponding to each user in the preliminary user group based on the rights, recommend the preliminary user rights to the corresponding users in the preliminary user group, and use the obtained rights redemption data as the preliminary recommendation result of the rights.

[0067] According to another embodiment of the present invention, the preliminary user group and the target user group are determined by using a reinforcement learning strategy.

[0068] According to another embodiment of the present invention, the initial decision tree model is constructed by a random forest model or a gradient stochastic model, and the constituent features of the initial decision tree model include: the user's basic features, the historical order features of the trading platform, and the historical browsing features of the trading platform.

[0069] According to another embodiment of the present invention, the rights recommendation device 500 further includes a result storage module (not shown in the figure), which is used to: after obtaining the recommendation result of the rights, persistently store the recommendation result of the rights if the recommendation result of the rights meets a preset result threshold.

[0070] According to another embodiment of the present invention, the benefit recommendation device 500 further includes a result optimization module (not shown in the figure), configured to: when the benefit recommendation result does not meet the result threshold, repeatedly execute the following steps until the benefit recommendation result meets the result threshold: construct a new training sample set based on the benefit recommendation result, train the decision tree model to obtain a decision tree optimization model; determine new user benefits corresponding to each user in the target user group based on the decision tree optimization model, recommend the corresponding new user benefits to each user in the target user group, and obtain a new benefit recommendation result.

[0071] Figure 6 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied.

[0072] like Figure 6 As shown, system architecture 600 may include terminal devices 601, 602, and 603, a network 604, and a server 605. Network 604 serves as the medium for providing communication links between terminal devices 601, 602, and 603 and server 605. Network 604 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0073] Users can use terminal devices 601, 602, and 603 to interact with server 605 via network 604 to receive or send messages, etc. Various communication client applications, such as benefit recommendation applications, can be installed on terminal devices 601, 602, and 603 (for example only).

[0074] Terminal devices 601, 602, and 603 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0075] Server 605 can be a server providing various services, such as a backend management server supporting the rights and interests exercised by users using terminal devices 601, 602, and 603 (for example only). Upon receiving a rights and interests recommendation request, the backend management server can determine a target user group and the corresponding user rights and interests for each user within that target user group. These user rights and interests are determined by a pre-built decision tree model, the training samples of which are derived from preliminary recommendation results of the rights and interests. The server can then recommend the corresponding user rights and interests to each user in the target user group, obtain the recommendation results, and feed back the processing results (e.g., the recommendation results of rights and interests – for example only) to the terminal devices.

[0076] It should be noted that the method for recommending rights provided in the embodiments of the present invention is generally executed by server 605, and correspondingly, the device for recommending rights is generally set in server 605.

[0077] It should be understood that Figure 6 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0078] The following is for reference. Figure 7 It shows a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers of the present invention. Figure 7 The terminal device or server shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0079] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 702 or programs loaded from storage section 708 into random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the system 700. The CPU 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0080] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, mouse, etc.; an output section 707 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 710 as needed so that computer programs read from it can be installed into the storage section 708 as needed.

[0081] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 709, and / or installed from removable medium 711. When the computer program is executed by central processing unit (CPU) 701, it performs the functions defined above in the system of this invention.

[0082] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0083] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0084] The units described in the embodiments of the present invention can be implemented in software or in hardware. The described units can also be housed in a processor; for example, a processor can be described as including a user rights determination module and a recommendation result determination module.

[0085] In some cases, the names of these modules do not constitute a limitation on the module itself. For example, the recommendation result determination module can also be described as "a module for recommending the corresponding user rights to each user in the target user group and obtaining the recommendation results of the rights".

[0086] On the other hand, the present invention also provides a computer-readable medium, which may be included in the device described in the embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to: in response to receiving a benefit recommendation request, determine a target user group and user benefits corresponding to each user in the target user group, wherein the user benefits are determined from the benefits by a pre-constructed decision tree model, the training samples of which are derived from preliminary recommendation results of the benefits; recommend the corresponding user benefits to each user in the target user group, and obtain the recommendation results of the benefits.

[0087] The technical solution of the present invention has the following advantages or beneficial effects: by responding to a received benefit recommendation request, determining the target user group and the user benefits corresponding to each user in the target user group, the user benefits are determined from the benefits by a pre-constructed decision tree model, and the training samples of the decision tree model come from the preliminary recommendation results of the benefits; the technical solution of recommending the corresponding user benefits to each user in the target user group and obtaining the benefit recommendation results realizes an efficient and low-cost benefit recommendation method. Compared with existing recommendation methods, it reduces the recommendation cost, and the user benefits corresponding to each user obtained based on the decision tree model ensure the rationality of the recommendation, improve the recommendation efficiency, increase the benefit redemption rate, and help predict subsequent benefit demand.

[0088] The specific embodiments described herein do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for recommending rights, characterized in that, include: In response to receiving a rights recommendation request, a target user group is determined, as well as the user rights corresponding to each user in the target user group. The user rights are determined from the rights by a pre-built decision tree model, and the training samples of the decision tree model come from the preliminary recommendation results of the rights. Recommend the corresponding user benefits to each user in the target user group, and obtain the recommendation results of the benefits.

2. The method according to claim 1, characterized in that, Before determining the user rights corresponding to each user in the target user group, the method further includes: Based on the preliminary recommendation results of the rights and benefits, a training sample set is constructed, and a preset decision tree initial model is trained to obtain a decision tree model. The decision tree model has the ability to predict the user rights and benefits corresponding to all users.

3. The method according to claim 2, characterized in that, Before constructing the training sample set based on the preliminary recommendation results of the aforementioned rights and interests, the method further includes: The target user group is determined based on the preset user selection rules; Based on the aforementioned rights, determine the user rights corresponding to each user in the user group being surveyed, and recommend the user rights to the corresponding users in the user group being surveyed. The obtained rights redemption data is used as the survey recommendation result of the rights.

4. The method according to claim 3, characterized in that, The preliminary user group and the target user group are determined using a reinforcement learning strategy.

5. The method according to claim 2, characterized in that, The initial decision tree model is constructed using a random forest model or a gradient stochastic model. The components of the initial decision tree model include: the user's basic characteristics, the historical order characteristics of the trading platform, and the historical browsing characteristics of the trading platform.

6. The method according to claim 1, characterized in that, After obtaining the recommendation results for the aforementioned rights, the method further includes: If the recommendation result of the right meets the preset result threshold, the recommendation result of the right is persistently stored.

7. The method according to claim 6, characterized in that, If the recommended benefit does not meet the result threshold, the method further includes: Repeat the following steps until the recommended result of the benefit meets the result threshold: Based on the recommendation results of the rights and interests, a new training sample set is constructed, the decision tree model is trained, and an optimized decision tree model is obtained. Based on the decision tree optimization model, new user benefits are determined for each user in the target user group, and the corresponding new user benefits are recommended to each user in the target user group to obtain new recommendation results for the benefits.

8. A rights recommendation device, characterized in that, include: The user rights determination module is used to respond to a received rights recommendation request, determine the target user group, and the user rights corresponding to each user in the target user group. The user rights are determined from the rights by a pre-built decision tree model, and the training samples of the decision tree model come from the preliminary recommendation results of the rights. The recommendation result determination module is used to recommend the corresponding user benefits to each user in the target user group and obtain the recommendation results of the benefits.

9. A mobile electronic device terminal, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images