Policy updating method and device

Through the policy update method, the problem of insufficient crowd selection ability in the EDM push strategy is solved, and accurate screening and recalling target business users is achieved, improving the effectiveness of push strategy and user reachability.

CN114676328BActive Publication Date: 2024-12-06HANGZHOU ALIBABA INT INTERNET IND CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202210323056.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-30
Publication Date
2024-12-06
Estimated Expiration
2042-03-30

AI Technical Summary

Technical Problem

The existing EDM (Email Direct Marketing) push strategy has weak expression ability and is difficult to expand when selecting crowd circles and pushing information, making it difficult to effectively determine the preferred population, affecting recall efficiency and unsubscribe rate.

Method used

Provide a policy update method, by receiving push policy update requests, determine the initial push parameters, push parameter step size and iteration cycle, calculate push crowd information, filter intermediate push parameters, use preset convergence functions to filter target push parameters, and update push policies To maximize the positive recall of business users.

Benefits of technology

It realizes accurate screening and recall of target business users, improves the effectiveness of push strategies, enhances the reachability of users with demand, and balances the recall efficiency and unsubscribe rate.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114676328B_ABST
    Figure CN114676328B_ABST
Patent Text Reader

Abstract

The embodiments of the present specification provide a policy updating method and device, wherein the policy updating method comprises: receiving a policy updating request submitted for a push policy for a target business; determining an initial push parameter, a push parameter step, and a plurality of push iteration cycles according to the policy updating request; calculating push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step, and determining a plurality of intermediate push parameters according to the push population information; filtering a target push parameter from the plurality of intermediate push parameters according to a preset convergence function; updating the push policy based on the target push parameter, and determining a set of business users associated with the target business using the updated push policy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present specification relate to the field of computer technology, and in particular to a policy updating method and device. Background Art

[0002] With the development of Internet technology, more and more businesses are going online, and in order to provide high-quality online services to users in need, large-scale promotion is usually used to reach effective users. EDM (Email Direct Marketing) is one of the important means. It actively pushes platform-related content or messages to users in the form of emails, and undertakes the important task of reaching effective users. The main implementation of EDM-based information push is based on crowd selection, and the size of the crowd directly determines the upper limit of the push success rate. However, if the crowd is expanded infinitely, negative indicators such as the unsubscription rate will be too high, which will also affect the health of the channel. Therefore, an effective solution is urgently needed to balance the recall efficiency and unsubscription rate. Summary of the invention

[0003] In view of this, an embodiment of this specification provides a policy updating method. One or more embodiments of this specification also relate to a policy updating device, a computing device, a computer-readable storage medium and a computer program to solve the technical defects existing in the prior art.

[0004] According to a first aspect of an embodiment of this specification, a policy updating method is provided, including:

[0005] Receive a policy update request submitted for a push policy of a target business;

[0006] Determining an initial push parameter, a push parameter step size, and a plurality of push iteration cycles according to the policy update request;

[0007] Calculating the push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step, and determining a plurality of intermediate push parameters according to the push population information;

[0008] Selecting a target push parameter from the plurality of intermediate push parameters according to a preset convergence function;

[0009] The push policy is updated based on the target push parameter, and the service user set associated with the target service is determined using the updated push policy.

[0010] According to a second aspect of an embodiment of this specification, a policy updating device is provided, including:

[0011] A receiving module, configured to receive a policy update request submitted for a push policy of a target service;

[0012] A determination module, configured to determine an initial push parameter, a push parameter step size, and a plurality of push iteration cycles according to the policy update request;

[0013] A calculation module is configured to calculate the push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step size, and determine a plurality of intermediate push parameters according to the push population information;

[0014] A screening module, configured to screen a target push parameter from the plurality of intermediate push parameters according to a preset convergence function;

[0015] The updating module is configured to update the push policy based on the target push parameter, and determine the service user set associated with the target service by using the updated push policy.

[0016] According to a third aspect of an embodiment of this specification, a computing device is provided, including:

[0017] Memory and processor;

[0018] The memory is used to store computer-executable instructions, and the processor is used to implement any step of the above-mentioned policy updating method when executing the computer-executable instructions.

[0019] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned policy updating method are implemented.

[0020] According to a fifth aspect of the embodiments of this specification, a computer program is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned policy updating method.

[0021] The strategy update method provided in this specification, in order to maximize the positive recall of effective users, can determine the initial push parameters, push parameter step size and multiple push iteration cycles according to the strategy update request after receiving the strategy update request submitted for the push strategy of the target business; then use the initial push parameters and push parameter step size to calculate the push population information corresponding to each push iteration cycle, so as to determine the preferred multiple intermediate push parameters according to the push population information, and in order to locate the reasonable push population of the target business, the target push parameters can be screened from the multiple intermediate push parameters according to the preset convergence function, and finally the push strategy can be updated based on the target push parameters, and the business user set associated with the target business can be determined using the updated push strategy, so that by calculating the push population information of each cycle, the target push parameters can be accurately screened out, so as to maximize the positive recall of business users through the parameters, so that the target business can reach users in need, so as to provide corresponding business services to these users. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a flow chart of a strategy updating method provided by an embodiment of this specification;

[0023] Figure 2 is a process flow chart of a strategy update method provided by an embodiment of this specification;

[0024] Figure 3 It is a structural diagram of a strategy updating device provided by an embodiment of this specification;

[0025] Figure 4 It is a structural block diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION

[0026] Many specific details are described in the following description to facilitate a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of this specification, so this specification is not limited to the specific implementation disclosed below.

[0027] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms of "a", "said" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0028] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0029] First, the terms involved in one or more embodiments of this specification are explained.

[0030] EDM (Email Direct Marketing): refers to actively pushing platform-related content or messages to users in the form of emails, that is, sending EDM emails to target customers, establishing communication channels with target customers, and directly conveying relevant information to them; sending electronic advertisements, product information, sales information, market research, marketing promotion activity information, etc.

[0031] Simulated annealing algorithm: It is derived from the principle of solid annealing and is a probability-based algorithm. The solid is heated to a sufficiently high temperature and then slowly cooled. When heated, the particles inside the solid become disordered as the temperature rises, and the internal energy increases. When slowly cooled, the particles gradually become ordered, reaching an equilibrium state at each temperature, and finally reaching the ground state at room temperature, and the internal energy is reduced to a minimum. In essence, it is an optimization algorithm that effectively avoids falling into a local minimum and eventually tends to the most reasonable serial structure globally by giving the search process a time-varying probability jump that eventually tends to zero.

[0032] In this specification, a policy updating method is provided. This specification also relates to a policy updating device, a computing device, a computer-readable storage medium and a computer program, which are described in detail one by one in the following embodiments.

[0033] In actual applications, when using EDM to target people and push information, the main idea is to assume that more active users are, the easier it is to open, and the more tolerant they are to EDM interruptions. However, this idea is mainly based on an ideal environment, and the biggest problem with this assumption is that it has a weak ability to describe the crowd, is inflexible, and difficult to expand, limiting the size of the crowd. It is difficult to effectively determine the preferred crowd corresponding to the EDM solution. Therefore, an effective solution is urgently needed to solve the above problems.

[0034] The strategy update method provided in this specification, in order to maximize the positive recall of effective users, can determine the initial push parameters, push parameter step size and multiple push iteration cycles according to the strategy update request after receiving the strategy update request submitted for the push strategy of the target business; then use the initial push parameters and push parameter step size to calculate the push population information corresponding to each push iteration cycle, so as to determine the preferred multiple intermediate push parameters according to the push population information, and in order to locate the reasonable push population of the target business, the target push parameters can be screened from the multiple intermediate push parameters according to the preset convergence function, and finally the push strategy can be updated based on the target push parameters, and the business user set associated with the target business can be determined using the updated push strategy, so that by calculating the push population information of each cycle, the target push parameters can be accurately screened out, so as to maximize the positive recall of business users through the parameters, so that the target business can reach users in need, so as to provide corresponding business services to these users.

[0035] Figure 1 A flow chart of a policy updating method provided according to an embodiment of the present specification is shown, which specifically includes the following steps.

[0036] Step S102: receiving a policy update request submitted for a push policy of a target service.

[0037] Specifically, the target business refers to the business that can be recommended through EDM, including but not limited to shopping business, game business, video business, application business, etc. In other words, it is necessary to use EDM to determine the business user cluster corresponding to the above business and recommend business-related information to it in order to achieve the purpose of promotion, while reaching more users who need business services and improving user experience. Correspondingly, the push strategy specifically refers to the strategy of determining the user cluster corresponding to the target business user and pushing information related to the target business to the user cluster. The strategy determines the size of the user cluster and the success rate of pushing messages. In other words, as long as the push strategy is set reasonably, it is possible to successfully reach more users in a larger user cluster, thereby achieving the purpose of effective promotion. Correspondingly, the strategy update request specifically refers to a request to update the push strategy associated with the target business. The request can be triggered when the push strategy fails to achieve the expected push effect, or it can be actively triggered by the business management party according to demand, so as to update a better push strategy and improve the exposure rate of the target business.

[0038] In actual applications, the triggering method of the policy update request can be set according to the actual application scenario, and this embodiment does not make any limitation here. Further, in order to facilitate the description of the policy update method, this embodiment takes the target business as the shopping business as an example, that is, it is necessary to update the push strategy that meets its expected promotion effect for the shopping platform, so as to reach more users in a larger population and achieve the purpose of effective promotion. The policy update methods in other scenarios can refer to the same or corresponding description content in this embodiment, and no more details will be given here.

[0039] Furthermore, in order to accurately update the push strategy, achieve more flexible and detailed positioning of effective user groups, and maximize positive recall of users, a preliminary relationship derivation can be performed before the strategy is updated to clarify the direction of strategy adjustment, so as to update the push strategy that meets business needs. In this embodiment, the specific implementation method is as follows:

[0040] Determine a set of subscribed users and a set of unsubscribed users corresponding to the target service, and build a service push relationship corresponding to the target service based on the set of subscribed users and the set of unsubscribed users; determine the initial push parameters associated with the target service based on the service push relationship; and upon receiving an update instruction submitted for the push policy, generate the policy update request corresponding to the push policy based on the policy update parameters carried in the update instruction and the initial push parameters.

[0041] Specifically, the subscription user set refers to the set of users who have successfully pushed messages during the historical message push process; that is, when information about the target business is sent to users by email, these users open the email and effectively read the information about the target business. Correspondingly, the unsubscribe user set refers to the set of users who have failed to push messages during the historical message push process; that is, when information about the target business is sent to users by email, these users do not open the email. Correspondingly, the business push relationship refers to the inherent relationship between the unsubscribe user set and the subscription user set, which can reflect the number of users that can be effectively reached by the target business. In other words, the business push relationship can characterize the user clusters that the target business urgently needs to push information. Correspondingly, the initial push parameter refers to the parameter for filtering the user cluster for the target business, that is, the initial push parameter is the filtering condition for filtering the user cluster associated with the target business. By adjusting the parameter, the number contained in the user cluster can be changed, thereby affecting the success rate of information push for the target business. Therefore, it is only necessary to accurately adjust the initial push parameter to complete the update of the push strategy.

[0042] Accordingly, the update instruction specifically refers to the instruction automatically triggered by the system or actively submitted by the management party of the target business, which is used to instruct the update of the recommended strategy; accordingly, the strategy update parameter specifically refers to the preset or manually set parameters required to update the recommended strategy. For example, if the most reasonable push parameters need to be solved, the solution cycle parameters and solution time parameters can be set, so as to achieve the purpose of determining the target push parameters and updating the push strategy.

[0043] Based on this, in order to accurately complete the policy update, the subscription user set and the unsubscription user set corresponding to the target business can be determined first, and the business push relationship corresponding to the target business can be determined based on the subscription user set and the unsubscription user set. At this time, the parameters that affect the success of the target business push can be reflected according to the business push relationship. Therefore, the initial push parameters associated with the target business can be determined using the business push relationship. Later, when an update instruction submitted for the push strategy is received, a policy update request can be generated based on the policy update parameters and initial push parameters carried in the update instruction, so as to realize the subsequent triggering of the update processing operation according to the policy update request to generate the target push parameters corresponding to the target business.

[0044] In actual applications, the update instruction can be controlled by the generation of the initial push parameters. When the target business uses the initial push parameters to determine the user cluster and push messages, if the push effect does not meet expectations, an update instruction can be automatically generated at this time, or a reminder can be sent to the business management party, so that the business management party can create an update instruction according to the needs, so that a policy update request that triggers the update policy can be generated later. In specific implementation, the triggering of the update instruction can be set according to the needs, and this embodiment does not make any limitation here.

[0045] In the specific implementation, in the process of creating a policy update request, in order to accurately calculate the target push parameters later, the user cluster determination scheme associated with the target business can be determined by the following formula (1):

[0046] isTarget(u)=isGood(u)-isBad(u) (1)

[0047] Among them, u represents the user, isTarget represents the business user set corresponding to the target business, isGood represents users who like emails and tend to open emails, that is, the subscription user set, and isBad represents users who do not like emails and tend to unsubscribe from emails, that is, the unsubscribe user set. It can be seen that the business user set corresponding to the target business should belong to users who like to open emails and will not unsubscribe from emails. Therefore, in order to effectively reach more users, it is necessary to accurately determine the subscription user set and the unsubscribe user set.

[0048] Based on this, the two user sets can be determined by the following formulas (2) and (3):

[0049] objective:max recallUV(u) (2)

[0050] stunsubscribeUV(u)≤threshold (3)

[0051] Among them, recallUV is the user who opened the email, that is, the corresponding subscription user set; the goal of formula (2) is to select the user set that satisfies formula (1) so that the number of users who opened the email is maximized. unsubscribeUV is the user who unsubscribes from the email, that is, the corresponding unsubscribe user set. The goal of formula (3) is to select the user set that satisfies formula (1) while making the number of users who unsubscribe from the email less than the given threshold, so as to ensure that the final business user set is more in line with the push strategy of the target business.

[0052] Based on the above analysis, we can know that as long as we consider the user activity related to the target business, we can determine the users that the information push of the target business can effectively reach. Therefore, the corresponding relationship between the target business and the initial push parameters can be established through the following formula (4):

[0053]

[0054] Among them, login(u, a i , b i ) is the true description of user u in the recent a i Days to visit target business at least b i times, i.e. a i Indicates the number of days of visit, b i Indicates the number of visits, i.e. a i and b i is the initial push parameter corresponding to the target business; therefore, as long as the number of days visited is small enough and the number of visits is large enough, the determined user cluster will be easier to reach. However, considering that the setting of these two parameters will reduce the number of users in the user cluster, it will also affect the push effect. Therefore, isBad needs to be used as a constraint condition to obtain a i and b i The most appropriate understanding.

[0055] Furthermore, if the recent user activity level does not meet the standard, business-related emails should not be pushed to users. The target group selection condition needs to maximize the number of recalled users under the condition that the number of users who have unsubscribed from the email is satisfied in order to achieve the expected promotion effect. Therefore, the solution for selecting the target group can increase the number of selected people by increasing the number of visit days a or reducing the number of visits b under the above premise. Therefore, the following formula (5) can be used to establish a model expression of the target group to determine a i and b i The most appropriate understanding is to select more user groups when the number of visit days and the number of visits are balanced and reasonable.

[0056]

[0057] Here, i represents the number of visits. That is, by solving the above formula (5) for a when i=1, 1 , a when i = 2 2 , and a when i = 3 3 , the optimal time lengths for visits 1, 2, and 3 can be obtained, and a sufficiently large set of business users can be selected according to the visit time and the number of visits.

[0058] For the convenience of description, this embodiment only takes the value of i to 3 as an example for explanation. In actual application scenarios, the value of i can be set according to the actual application scenarios, and this embodiment does not make any limitation here.

[0059] Step S104: determining an initial push parameter, a push parameter step size, and a plurality of push iteration cycles according to the policy update request.

[0060] Specifically, after determining the policy update request as mentioned above, further, considering that the policy update request needs to update the push policy of the target business, and the push policy is controlled by the initial push parameters, the initial push parameters determine the number of days and times of user visits, so the initial push parameters will determine the size of the business user set corresponding to the target business. In order to maximize the recall of users, the initial push parameters, push parameter step size and multiple push iteration cycles can be determined according to the policy update request, so as to facilitate the subsequent solution of the most reasonable push parameters in combination with the simulated annealing algorithm, so as to update the push policy to a strategy that meets business expectations.

[0061] Among them, the initial push parameter specifically refers to the parameter for screening user clusters for the target business, that is, the initial push parameter is the screening condition for screening user clusters associated with the target business. By adjusting the parameter, the number contained in the user cluster can be changed, thereby affecting the information push success rate of the target business. Therefore, it is only necessary to accurately adjust the initial push parameter to complete the update of the push strategy. Correspondingly, the push parameter step size specifically refers to the parameter step size corresponding to each step cycle when calculating the target push parameter based on the simulated annealing algorithm, which is used to accurately calculate the number of visit days. Correspondingly, multiple push iteration cycles specifically refer to the number of cycles that need to be iteratively calculated in each step cycle when calculating the target push parameter based on the simulated annealing algorithm. It can be set according to actual needs, and this embodiment does not make any limitation here.

[0062] Furthermore, when determining the initial push parameters, the push parameter step size, and multiple push iteration cycles, it is considered that the parameters that determine the push strategy and have an impact on the business user set are the number of days visited and the number of visits, and the number of days visited and the number of visits are solved by the simulated annealing algorithm. Therefore, it is necessary to determine the relevant parameters by reading. In this embodiment, the specific implementation method is as follows:

[0063] Determine a policy update rule according to the policy update request, and read a push time parameter and a push number parameter according to the policy update rule to form the initial push parameter; parse the policy update request to obtain the push parameter step and the multiple push iteration cycles.

[0064] Specifically, the policy update rule specifically refers to the rule for updating the initial push parameters, and the initial push parameters are related to time and number of times. Therefore, the push time parameter and push number parameter associated with the target service can be read based on the policy update rule. Among them, the push time parameter specifically refers to the number of days of access, and the push number parameter specifically refers to the number of accesses. The push time parameter and the push number parameter are associated with each other, and together determine the number of users in the service user set.

[0065] Based on this, after determining the policy update rule according to the policy update request, the push time parameter and the push number parameter can be read according to the policy rule to form the initial push parameter; at the same time, the policy update request is parsed to obtain the push parameter step size and the multiple push iteration cycles. This enables the target push parameter to be solved based on the simulated annealing algorithm in the future, thereby achieving the purpose of updating the policy.

[0066] In summary, by adopting the strategy update rule to determine the parameters, it can be ensured that the target push parameters can be accurately calculated based on the simulated annealing algorithm in the future, thereby maximizing the positive recall of users.

[0067] Step S106, calculating the push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step size, and determining a plurality of intermediate push parameters according to the push population information.

[0068] Step S108: Filter the target push parameter from the plurality of intermediate push parameters according to a preset convergence function.

[0069] Specifically, after determining the initial push parameters, push parameter step size and multiple push iteration cycles as described above, further, in order to be able to characterize the user set corresponding to the push parameters in each cycle, the push population information corresponding to each push iteration cycle can be calculated based on the initial push parameters and the push parameter step size, so as to determine the target user cluster that can be reached by different push parameters in each push iteration cycle through a one-to-one calculation method, and finally determine multiple intermediate push parameters based on the push population information, which can be used to subsequently determine the target push parameters through a convergence method.

[0070] Among them, the push population information specifically refers to the number of target populations calculated using the initial push parameters in each push iteration cycle, and this number of populations is used to reflect the number of populations that can be circled by the push parameters. Correspondingly, the intermediate push parameters specifically refer to the push parameters that circle the largest number of people in each cycle. Finally, the target push parameters can be determined by comparing the push parameters corresponding to all cycles and converging. Correspondingly, the preset convergence function specifically refers to a function that selects a preset condition from multiple intermediate push parameters, and the functions involved are used to constrain the balance of push parameters on the number of user populations; correspondingly, the target push parameters specifically refer to the ability to reach enough business users based on the push parameters, that is, the target push parameters can maximize the positive recall of users under the limitation of limited push parameters.

[0071] On this basis, when login(u, a i , b i ) is used to express the business user clusters that need to be determined. In fact, different parameters will affect the number of user clusters:

[0072] (1) if b ≥ c, Under this condition, a represents the number of days of visit, and b and c both represent the number of visits; that is, if the number of days of visit is the same, the lower the number of visits, the larger the population covered, and conversely, the higher the number of visits, the smaller the population covered.

[0073] (2) if a≤c, Under this condition, a and c represent the number of visit days, and b represents the number of visits; that is, if the number of visits is the same, the greater the number of visit days, the larger the population covered, and conversely, the smaller the number of visit days, the smaller the population covered.

[0074] (3) if a≤c and b≥d, Under this condition, a and c represent the number of visit days, and b and d represent the number of visits; that is, the lower the number of visits and the greater the number of visit days, the larger the population covered; the higher the number of visits and the smaller the number of visit days, the smaller the population covered.

[0075] According to the above condition (2), there are multiple logins (u, a i , b) When the population is combined, only the largest a is retained i The maximum and reasonable business user set corresponding to the target business can be determined under a limited number of accesses. At the same time, according to condition (3), the user with the most accesses a i (Number of days visited) must be greater than the number of visits with the least number of a i (Visit days). That is, when the number of visits is 1, 2, and 3, a 1 ≤a 2 ≤a 3 For example, the target business needs to be expanded based on the number of people who visit at least 3 times in 180 days. To increase the number of people, the purpose of expansion can be achieved by increasing the number of visit days to 180 or reducing the number of visits to 3.

[0076] Based on this, in order to solve the most reasonable number of days for access when the number of accesses is fixed, and to combine the two to circle the maximum set of business users corresponding to the target business, the number of accesses can be determined first, and then the number of users who subscribe to emails can be used as a condition, and the number of users who unsubscribe from emails can be used as a constraint condition to calculate the most reasonable number of days for access. For the convenience of description, this embodiment uses the number of accesses of 1, 2, and 3 as examples for explanation. The calculation process of other number of accesses can refer to the same or corresponding description content of this embodiment, and no limitation is made here.

[0077] Furthermore, the relationship between the number of visit days and the number of user groups can be determined based on the above conditions, namely objective:max f(a 1 , a 2 , a 3 ); stg(a 1 , a 2 , a 3 )≤threshold. 1 is the number of days visited when the number of visits is 1; a 2 is the number of days visited when the number of visits is 2; a 3 is the number of visit days when the number of visits is 3.

[0078] Based on this, by calculating the most reasonable parameter combination {a1 , a 2 , a 3}, unsubscribe user g(a 1 , a 2 , a 3 ) satisfies the threshold condition, so that the recalled user UVf(a 1 , a 2 , a 3 ) is maximized, so that the maximum business user set corresponding to different access times can be determined; among them, the f function and the g function are both stable statistical functions, and the maximum business user set can be obtained by maximizing the constraint g through parameter combination.

[0079] On this basis, we only need to calculate a 1 , a 2 , a 3 The most appropriate solution corresponding to each of them can be obtained. On this basis, the simulated annealing algorithm can be used to solve it, that is, the parameter combination {a 1 , a 2 , a 3} as the numerator in the algorithm, push the parameter step size as the initial temperature in the algorithm, calculate the change of internal energy, and calculate the minimum value of internal energy at each temperature to get the most reasonable solution {a 1 , a 2 , a 3}, that is, by calculating the number of push groups within each step, the target push parameter {a 1 , a 2 , a 3 That is to say, by solving the maximum f of the constraint function g, we can get the target push parameter that can determine the maximum user set, that is, objective: max r(a 1 , a 2 , a 3 )=f(a 1 , a 2 , a 3 )-K*max[0.g(a 1 , a 2 , a 3 )-threshold], where K represents a hyperparameter, which can be set to a larger value to achieve the goal of unsubscribing users g(a 1 , a 2 , a 3 ) When it exceeds the threshold, it can act as a fuse.

[0080] Furthermore, in the simulated annealing algorithm, the parameter combination {a 1 , a 2 , a3}, you can first set a default access day for each access number, and then iterate the calculation based on this technology. For example, when the access number is 1, the access day a 1 The default value is 90. For example, if the number of visits is 2, the number of days of visit is a 2 The default value is 180. For example, if the number of visits is 3, the number of days of visit is a 3 The default value is 180, and calculations can be performed based on this.

[0081] Furthermore, the push iteration cycle can be set to N, and the push parameter step size can be set to α, which can be 5, 10, or 15. At the same time, in order to avoid the impact of simply increasing the step size on the accuracy, e can be set i As the weight, where e i ∈[-1,0,1], that is, when determining the number of days visited in the next push iteration cycle, a value can be randomly selected in [-1,0,1], and then multiplied by the push parameter step length α, and the product is added to the number of days visited in the previous push iteration cycle, and the number of days visited in the next push iteration cycle can be obtained, and the information of the push population can be calculated based on this. in, Indicates the number of days visited once in the next push iteration cycle. Indicates the number of days visited twice in the next push iteration cycle. Indicates the number of days visited 3 times in the next push iteration cycle. Then, the number of days visited and the number of visits are used to calculate the user cluster set corresponding to each push iteration cycle.

[0082] In this process, you need to set △ E Indicates the difference in the number of people pushed to the current push iteration cycle and the previous push iteration cycle, r n Indicates the number of people being pushed to in the current push iteration cycle, r n-1 Indicates the number of people pushed to in the previous push iteration cycle, and in △ E >0 or exp(-C△ E / α)(where C is the exploration parameter, and the number of access days corresponding to each push iteration cycle is updated, that is, a 1 Updated to a 2 Updated to a 3 Updated to Use it as the number of visit days for the next push iteration cycle and then perform further calculations.

[0083] By repeating the above calculation process and cooling down according to the push parameter step size e.gα=0.99α until the push iteration cycle N or the result converges, the most reasonable parameter combination {a 1 , a 2 , a 3}, then the business can determine the business user set based on this parameter.

[0084] Furthermore, when the target push parameters are calculated based on the simulated annealing algorithm, the push population information corresponding to each push iteration cycle is actually solved by iterative calculation, and then the intermediate push parameters that initially meet the conditions are selected according to the push population information, which are used for subsequent update strategies to determine the business user set. In this embodiment, the specific implementation method is as follows:

[0085] Use the initial push parameters as the benchmark push parameters corresponding to the j-th push iteration cycle; calculate the push population information corresponding to the j-th push iteration cycle according to the benchmark push parameters; determine whether the push population information corresponding to the j-th push iteration cycle meets the cycle end condition; if not, determine the benchmark push parameters corresponding to the j+1 push iteration cycle according to the push parameter step and the benchmark push parameters corresponding to the j-th push iteration cycle, j is incremented by 1, and execute the step of calculating the push population information corresponding to the j-th push iteration cycle according to the benchmark push parameters; if so, execute the step of determining multiple intermediate push parameters according to the push population information.

[0086] Specifically, the benchmark push parameters refer to the parameters used in the push iteration cycle to calculate the push population information, that is, the corresponding access days and access times; correspondingly, the cycle end condition refers to the condition that the maximum iteration cycle is reached or the number of push populations cannot be increased. Wherein, j is a positive integer, and j starts from 1.

[0087] Based on this, first determine the benchmark push parameters corresponding to the j-th push iteration cycle according to the initial push parameters; then use the benchmark push parameters to calculate the push population information corresponding to the j-th push iteration cycle; and determine whether the push population information corresponding to the j-th push iteration cycle meets the cycle end condition; if it does, it means that the push population information calculation of each push iteration cycle is completed, and the push population information can be used to determine the intermediate push parameters for subsequent screening of target push parameters. If it does not meet the requirements, it means that the push population information corresponding to the next push iteration cycle needs to be calculated, and the benchmark push parameters corresponding to the j+1 push iteration cycle are determined based on the benchmark push parameters of the j push iteration cycle. Therefore, it is necessary to first calculate the benchmark push parameters corresponding to the j+1 push iteration cycle based on the push parameter step size and the benchmark push parameters of the j push iteration cycle, and then j is incremented by 1, and the push population information can be determined until the conditions are met.

[0088] Furthermore, considering that multiple step sizes may be set in actual application scenarios to solve the target push parameters based on different time precisions, the reference push parameters will also be different under different step sizes. In this embodiment, the specific implementation method is as follows:

[0089] Determine the benchmark push parameter step corresponding to the i-th step cycle in the push parameter step; use the initial push parameter as the benchmark push parameter corresponding to the j-th push iteration cycle in the benchmark push parameter step corresponding to the i-th step cycle; accordingly, before executing the step of determining multiple intermediate push parameters according to the push population information, it also includes: judging whether there are remaining push parameter steps in the push parameter step; if so, determine the benchmark push parameter step corresponding to the i+1-th step cycle in the remaining push parameter step, i is incremented by 1, and execute the step of using the initial push parameter as the benchmark push parameter corresponding to the j-th push iteration cycle in the benchmark push parameter step corresponding to the i-th step cycle; if not, execute the step of determining multiple intermediate push parameters according to the push population information. Wherein, i starts at 1 and i is a positive integer.

[0090] Based on this, first determine the benchmark push parameter step corresponding to the i-th step cycle in the push parameter step, and then use the initial push parameter as the benchmark push parameter corresponding to the j-th push iteration cycle in the benchmark push parameter step corresponding to the i-th step cycle; then calculate the push population information corresponding to the j-th push iteration cycle according to the benchmark push parameter; and judge whether the push population information corresponding to the j-th push iteration cycle meets the cycle end condition; if not, it means that the push population information corresponding to the next push iteration cycle needs to be calculated, and the benchmark push parameter corresponding to the j+1 push iteration cycle is determined according to the benchmark push parameter of the j push iteration cycle. Therefore, it is necessary to first calculate the benchmark push parameter corresponding to the j+1 push iteration cycle according to the benchmark push parameter step corresponding to the i-th step cycle, and then j is incremented by 1, and the determination of the push population information can be continued; if so, it means that the calculation of the push population information of each push iteration cycle is completed, then it means that the calculation of the i-th step cycle is completed, and it can be judged again whether there are remaining push parameter steps in the push parameter step.

[0091] If it exists, it means that there is a benchmark push parameter step in the push parameter step that is different from the i-th step period, and the calculation needs to continue. At this time, the benchmark push parameter step corresponding to the i+1-th step period can be determined in the remaining push parameter step. After i is increased by 1, the process of iteratively calculating the push population information within the i+1 step period is executed; if it does not exist, it means that the calculation has been completed, and all the push population information can be integrated to determine the intermediate push parameters.

[0092] For example, in order to effectively reach more users on the shopping platform, EDM will be used to push messages. In this process, to ensure that the maximum number of people meets the expected demand, a1 is set to the number of days to access the email once, a2 is set to the number of days to access the email twice, and a3 is set to the number of days to access the email three times; at the same time, the iteration cycle is set to N, and the step length is 5, 10 and 15. The default number of days for a1 is set to 90, the default number of days for a2 is set to 180, and the default number of days for a3 is set to 180 for iterative calculation.

[0093] Further, the most reasonable values ​​of a1, a2 and a3 corresponding to the cycle with a step size of 5 are calculated, that is, a1=the number of users who visited the associated shopping platform to push emails once within 90 days is calculated as S511, a2=the number of users who visited the associated shopping platform to push emails twice within 180 days is calculated as S512, and a3=the number of users who visited the associated shopping platform to push emails 3 times within 180 days is calculated as S513; at this time, it is determined that the iteration cycle N has not been reached, and the number of visit days for the next iteration cycle is calculated according to the step size of 5 and a1=90, a2=180 and a3=180, and a1=95, a2=185 and a3=185 in the next cycle are determined, and so on, until the number of iterations in the current step size cycle reaches N, then the most reasonable values ​​of a1, a2 and a3 corresponding to the cycle with a step size of 10 are calculated.

[0094] Furthermore, after the values ​​corresponding to a1, a2 and a3 in each step cycle are calculated, the number of user groups corresponding to different values ​​of a1, a2 and a3 will be determined. On this basis, the most reasonable parameter combination set will be selected from multiple a1, a2 and a3 based on the number of user groups, so as to facilitate the subsequent identification of business user clusters according to the parameters.

[0095] In summary, by using the simulated annealing algorithm to calculate the target push parameters, the calculation and processing operations can be completed more lightly. At the same time, it is more interpretable, so as to achieve the purpose of accurately updating the push strategy.

[0096] In addition, since the push population information corresponds to each push iteration cycle, different push population information corresponding to different push parameters will be obtained, and the step cycle contains at least one, so it is necessary to first select the intermediate push parameters based on the step cycle. In this embodiment, the specific implementation method is as follows:

[0097] According to the push population information corresponding to each push iteration cycle, the target push population information corresponding to each step cycle is screened; according to the target push population information corresponding to each step cycle, the intermediate push parameters corresponding to each step cycle are determined; by integrating the intermediate push parameters corresponding to each step cycle, the multiple intermediate push parameters are obtained.

[0098] Specifically, the target push population information specifically refers to the population information with the largest number of people in each push iteration cycle. Based on this, after obtaining the push population information corresponding to each push iteration cycle, the maximum number of push populations corresponding to each step cycle can be selected as the target push population information, and then the push parameters calculated for the target push population information are selected as the intermediate push parameters, and then integrated into multiple intermediate push parameters.

[0099] Using the above example, when the step length is 5, the number of people pushed to each iteration cycle is {S511, S512, S513} {S521, S522, S523}... {S5N1, S5N2, S5N3}, and when the step length is 10, the number of people pushed to each iteration cycle is...; by comparing the number of people pushed to each step cycle, it is determined that when the step length is 5, the largest number of people pushed to is {S521, S522, S523}, when the step length is 10, the largest number of people pushed to is {S1011, S1012, S1013}, and when the step length is 15, the number of people pushed to is The largest number is {S1531, S1532, S1533}, and then determine that the number of visit days corresponding to {S521, S522, S523} are a1=100, a2=190, a3=190 respectively; similarly, determine that the number of visit days corresponding to {S1011, S1012, S1013} are a1=100, a2=190, a3=190 respectively; determine that the number of visit days corresponding to {S1531, S1532, S1533} are a1=135, a2=225, a3=225 respectively; use them as an intermediate parameter combination for subsequent use.

[0100] In summary, by determining the intermediate push parameters by comparing within a period, the intermediate push parameters can be determined more accurately, thereby facilitating the subsequent identification of effective users based on this.

[0101] Furthermore, the filtering of the target push parameters from the multiple intermediate push parameters according to the preset convergence function includes: determining the initial business user set corresponding to each intermediate push parameter according to the preset convergence function; calculating the unsubscription rate of the initial business user set according to the number of subscribed users and the number of unsubscribed users included in the initial business user set; determining the target business user set from the initial business user set corresponding to each intermediate push parameter according to the unsubscription rate, and using the intermediate push parameter corresponding to the target business user set as the target push parameter.

[0102] Specifically, the initial business user set specifically refers to the business user set corresponding to each intermediate push parameter, and the initial business user set includes subscribed users and unsubscribed users. The unsubscription rate obtained by the parameter can be calculated by combining the number of subscribed users and the number of unsubscribed users. If the unsubscription rate is greater than the threshold, it means that effective business users cannot be identified, so the intermediate push parameter with the lowest unsubscription rate can be selected as the target push parameter. That is, according to the unsubscription rate, the target business user set can be determined in the initial business user set corresponding to each intermediate push parameter, and the intermediate push parameter corresponding to the target business user set can be used as the target push parameter.

[0103] Continuing with the above example, after determining the intermediate visit days corresponding to each step {a1=100, a2=190, a3=190}{a1=100, a2=190, a3=190} and {a1=135, a2=225, a3=225}, the unsubscribe rate can be calculated based on the number of push groups corresponding to each group of visit days, and compared with the constraints. Finally, the most reasonable visit days can be selected as the target number of visits. According to the comparison results, {a1=100, a2=190, a3=190} is determined as the target visit days for subsequent determination of business user clusters and message push.

[0104] In summary, by adopting a preset convergence function to determine the target push parameters, it is possible to select the parameters that maximize the recall of users from multiple intermediate push parameters, and use this to perform subsequent strategy updates, so as to target a wider set of users for the target business.

[0105] Step S110: updating the push policy based on the target push parameter, and determining a set of service users associated with the target service using the updated push policy.

[0106] Specifically, after determining the target push parameters as mentioned above, the push strategy can be further updated according to the target push parameters so that the push strategy can be updated to a strategy that responds to the policy update request, and the updated push strategy can be used to define the set of business users associated with the target business, so that the target business can be effectively promoted and better serve the users in the business user set.

[0107] Further, the updating of the push policy based on the target push parameter and determining the set of service users associated with the target service by using the updated push policy includes:

[0108] Determine the service push parameters in the push strategy; update the service push parameters according to the target push parameters, and determine the target push strategy according to the update result; determine the target access time and the target access number according to the target push strategy, and determine the service user set associated with the target service according to the target access time and the target access number.

[0109] Specifically, the business push parameters specifically refer to the push parameters that have not been updated in the inference strategy, and the user set defined by the parameters cannot meet the current needs; correspondingly, the target push strategy specifically refers to the strategy of updating the business push parameters to the target push parameters; correspondingly, the target access time specifically refers to the maximum number of access days for clicking on the email; correspondingly, the target number of visits specifically refers to the minimum number of visits for clicking on the email corresponding to the target access days.

[0110] Based on this, in order to accurately update the push strategy, the business push parameters can be determined in the push strategy; then the business push parameters are updated according to the target push parameters to replace the business push parameters in the policy, thereby obtaining the target push strategy; when the business user set is circled, the target access time and target access number can be determined according to the target push strategy, and then the business user set associated with the target business can be determined according to the target access time and the target access number.

[0111] In summary, by using target push parameters to determine the target access time and target access times, and using this to target users, it is possible to more accurately target user sets that meet target business needs, and ensure that the users included in the business user set are more in line with expected results.

[0112] Furthermore, after determining the target service user set, information can be sent to each user, that is, the push information corresponding to the target service is determined; and the push information is sent to the service users included in the service user set according to the updated push policy. The push information specifically refers to the information associated with the target service, which is sent to each user in each service user set in the form of an email.

[0113] The strategy update method provided in this specification, in order to maximize the positive recall of effective users, can determine the initial push parameters, push parameter step size and multiple push iteration cycles according to the strategy update request after receiving the strategy update request submitted for the push strategy of the target business; then use the initial push parameters and push parameter step size to calculate the push population information corresponding to each push iteration cycle, so as to determine the preferred multiple intermediate push parameters according to the push population information, and in order to locate the reasonable push population of the target business, the target push parameters can be screened from the multiple intermediate push parameters according to the preset convergence function, and finally the push strategy can be updated based on the target push parameters, and the business user set associated with the target business can be determined using the updated push strategy, so that by calculating the push population information of each cycle, the target push parameters can be accurately screened out, so as to maximize the positive recall of business users through the parameters, so that the target business can reach users in need, so as to provide corresponding business services to these users.

[0114] The following combination Figure 2 , taking the application of the strategy update method provided in this specification in the game platform information push scenario as an example, the strategy update method is further explained. Figure 2 A processing flow chart of a policy updating method provided by an embodiment of the present specification is shown, which specifically includes the following steps.

[0115] Step S202: receiving a policy update request submitted for a push policy of a target service.

[0116] Step S204: determining a policy update rule according to the policy update request, and reading a push time parameter and a push number parameter according to the policy update rule to form the initial push parameter.

[0117] Step S206: parse the policy update request to obtain the push parameter step size and the multiple push iteration cycles.

[0118] Step S208: determining a reference push parameter step corresponding to the i-th step period in the push parameter step.

[0119] Step S210: In the reference push parameter step corresponding to the i-th step period, the initial push parameter is used as the reference push parameter corresponding to the j-th push iteration period.

[0120] Step S212: Calculate the push population information corresponding to the j-th push iteration cycle according to the benchmark push parameters.

[0121] Step S214, determine whether the push population information corresponding to the j-th push iteration cycle meets the cycle end condition; if not, execute step S216, if so, execute step S218.

[0122] Step S216, determine the benchmark push parameter corresponding to the j+1 push iteration period according to the benchmark push parameter step corresponding to the i-th step period and the benchmark push parameter corresponding to the j-th push iteration period, j is incremented by 1, and step S212 is executed.

[0123] Step S218, determining whether there are any remaining push parameter steps in the push parameter step; if so, executing step S220; if not, executing step S222.

[0124] Step S220, determining the reference push parameter step corresponding to the (i+1)th step period from the remaining push parameter step, i is incremented by 1, and executing step S210.

[0125] Step S222, according to the push population information corresponding to each push iteration cycle, the target push population information corresponding to each step cycle is filtered.

[0126] Step S224, determining the intermediate push parameters corresponding to each step period according to the target push population information corresponding to each step period.

[0127] Step S226: obtaining the plurality of intermediate push parameters by integrating the intermediate push parameters corresponding to each step period.

[0128] Step S228: determining an initial service user set corresponding to each intermediate push parameter according to the preset convergence function.

[0129] Step S230: calculating the unsubscription rate of the initial service user set according to the number of subscribed users and the number of unsubscribed users included in the initial service user set.

[0130] Step S232: determining a target service user set from the initial service user set corresponding to each intermediate push parameter according to the unsubscription rate, and using the intermediate push parameter corresponding to the target service user set as the target push parameter.

[0131] Step S234, determining a service push parameter in the push strategy, updating the service push parameter according to the target push parameter, and determining a target push strategy according to the update result.

[0132] Step S236: determine the target access time and the target access number according to the target push strategy, and determine the service user set associated with the target service according to the target access time and the target access number.

[0133] Further, push information corresponding to the target service is determined; and the push information is sent to the service users included in the service user set according to the updated push policy.

[0134] To summarize, in order to maximize the positive recall of effective users, after receiving the policy update request submitted for the push strategy of the target business, the initial push parameters, push parameter step size and multiple push iteration cycles can be determined according to the policy update request; then, the push population information corresponding to each push iteration cycle is calculated using the initial push parameters and the push parameter step size, so as to determine the preferred multiple intermediate push parameters according to the push population information, and in order to locate the reasonable push population for the target business, the target push parameters can be screened from the multiple intermediate push parameters according to the preset convergence function, and finally the push strategy can be updated based on the target push parameters, and the business user set associated with the target business can be determined using the updated push strategy. By calculating the push population information of each cycle, the target push parameters can be accurately screened out, so as to maximize the positive recall of business users through the parameters, so that the target business can reach users in need and provide corresponding business services to these users.

[0135] Corresponding to the above method embodiment, this specification also provides a strategy updating device embodiment, Figure 3 FIG. 2 shows a schematic diagram of a strategy updating device provided by an embodiment of the present specification. Figure 3 As shown, the device comprises:

[0136] The receiving module 302 is configured to receive a policy update request submitted for a push policy of a target service;

[0137] A determination module 304 is configured to determine an initial push parameter, a push parameter step size, and a plurality of push iteration cycles according to the policy update request;

[0138] The calculation module 306 is configured to calculate the push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step size, and determine a plurality of intermediate push parameters according to the push population information;

[0139] A screening module 308 is configured to screen a target push parameter from the plurality of intermediate push parameters according to a preset convergence function;

[0140] The updating module 310 is configured to update the push policy based on the target push parameter, and determine the service user set associated with the target service by using the updated push policy.

[0141] In an optional embodiment, the receiving module 302 is further configured to:

[0142] Determine a set of subscribed users and a set of unsubscribed users corresponding to the target service, and build a service push relationship corresponding to the target service based on the set of subscribed users and the set of unsubscribed users; determine the initial push parameters associated with the target service based on the service push relationship; and upon receiving an update instruction submitted for the push policy, generate the policy update request corresponding to the push policy based on the policy update parameters carried in the update instruction and the initial push parameters.

[0143] In an optional embodiment, the determining module 304 is further configured to:

[0144] Determine a policy update rule according to the policy update request, and read a push time parameter and a push number parameter according to the policy update rule to form the initial push parameter; parse the policy update request to obtain the push parameter step and the multiple push iteration cycles.

[0145] In an optional embodiment, the calculation module 306 is further configured to:

[0146] Use the initial push parameters as the benchmark push parameters corresponding to the j-th push iteration cycle; calculate the push population information corresponding to the j-th push iteration cycle according to the benchmark push parameters; determine whether the push population information corresponding to the j-th push iteration cycle meets the cycle end condition; if not, determine the benchmark push parameters corresponding to the j+1 push iteration cycle according to the push parameter step and the benchmark push parameters corresponding to the j-th push iteration cycle, j is incremented by 1, and execute the step of calculating the push population information corresponding to the j-th push iteration cycle according to the benchmark push parameters; if so, execute the step of determining multiple intermediate push parameters according to the push population information.

[0147] In an optional embodiment, the calculation module 306 is further configured to:

[0148] Determine the benchmark push parameter step corresponding to the i-th step period in the push parameter step; use the initial push parameter as the benchmark push parameter corresponding to the j-th push iteration period in the benchmark push parameter step corresponding to the i-th step period; accordingly, before executing the step of determining multiple intermediate push parameters according to the push population information, it also includes: judging whether there are remaining push parameter steps in the push parameter step; if so, determine the benchmark push parameter step corresponding to the i+1-th step period in the remaining push parameter step, i is incremented by 1, and execute the step of using the initial push parameter as the benchmark push parameter corresponding to the j-th push iteration period in the benchmark push parameter step corresponding to the i-th step period; if not, execute the step of determining multiple intermediate push parameters according to the push population information.

[0149] In an optional embodiment, the calculation module 306 is further configured to:

[0150] According to the push population information corresponding to each push iteration cycle, the target push population information corresponding to each step cycle is screened; according to the target push population information corresponding to each step cycle, the intermediate push parameters corresponding to each step cycle are determined; by integrating the intermediate push parameters corresponding to each step cycle, the multiple intermediate push parameters are obtained.

[0151] In an optional embodiment, the screening module 308 is further configured to:

[0152] Determine the initial business user set corresponding to each intermediate push parameter according to the preset convergence function; calculate the unsubscription rate of the initial business user set according to the number of subscribed users and the number of unsubscribed users contained in the initial business user set; determine the target business user set in the initial business user set corresponding to each intermediate push parameter according to the unsubscription rate, and use the intermediate push parameter corresponding to the target business user set as the target push parameter.

[0153] In an optional embodiment, the updating module 310 is further configured to:

[0154] Determine the service push parameters in the push strategy; update the service push parameters according to the target push parameters, and determine the target push strategy according to the update result; determine the target access time and the target access number according to the target push strategy, and determine the service user set associated with the target service according to the target access time and the target access number.

[0155] In an optional embodiment, the policy updating device further includes:

[0156] The push information determination module is configured to determine the push information corresponding to the target service; and send the push information to the service users included in the service user set according to the updated push strategy.

[0157] The strategy updating device provided in the present specification can, in order to maximize the positive recall of valid users, determine the initial push parameters, push parameter step size and multiple push iteration cycles according to the strategy update request after receiving the strategy update request submitted for the push strategy of the target business; then use the initial push parameters and push parameter step size to calculate the push population information corresponding to each push iteration cycle, so as to determine the preferred multiple intermediate push parameters according to the push population information; and in order to locate the reasonable push population of the target business, the target push parameters can be screened from the multiple intermediate push parameters according to the preset convergence function; finally, the push strategy can be updated based on the target push parameters, and the business user set associated with the target business can be determined using the updated push strategy, so as to achieve that by calculating the push population information of each cycle, the target push parameters can be accurately screened out, so as to maximize the positive recall of business users through the parameters, so that the target business can reach users in need, so as to provide corresponding business services to these users.

[0158] The above is a schematic scheme of a policy updating device of this embodiment. It should be noted that the technical scheme of the policy updating device and the technical scheme of the above policy updating method belong to the same concept, and the details not described in detail in the technical scheme of the policy updating device can be referred to the description of the technical scheme of the above policy updating method.

[0159] Figure 4 The block diagram of a computing device 400 according to one embodiment of the present specification is shown. The components of the computing device 400 include but are not limited to a memory 410 and a processor 420. The processor 420 is connected to the memory 410 via a bus 430, and the database 450 is used to store data.

[0160] The computing device 400 also includes an access device 440 that enables the computing device 400 to communicate via one or more networks 460. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 440 may include one or more of any type of network interface (e.g., a network interface card (NIC)) wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a World Wide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.

[0161] In one embodiment of the present specification, the above components of the computing device 400 and Figure 4 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 4The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0162] Computing device 400 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or PC. Computing device 400 may also be a mobile or stationary server.

[0163] The processor 420 is used to execute the following computer executable instructions, which implement the steps of the above-mentioned policy updating method when executed by the processor.

[0164] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above-mentioned policy update method belong to the same concept, and the details of the technical scheme of the computing device not described in detail can all be referred to the description of the technical scheme of the above-mentioned policy update method.

[0165] An embodiment of the present specification further provides a computer-readable storage medium storing computer-executable instructions, which can implement the steps of the above-mentioned policy updating method when executed by a processor.

[0166] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the above-mentioned policy update method belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be referred to the description of the technical scheme of the above-mentioned policy update method.

[0167] An embodiment of the present specification further provides a computer program, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned policy updating method.

[0168] The above is an illustrative solution of a computer program of this embodiment. It should be noted that the technical solution of the computer program and the technical solution of the above-mentioned policy update method belong to the same concept, and the details not described in detail in the technical solution of the computer program can be referred to the description of the technical solution of the above-mentioned policy update method.

[0169] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0170] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0171] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of this specification are not limited by the order of the actions described, because according to the embodiments of this specification, some 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 preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of this specification.

[0172] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0173] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The optional embodiments do not describe all the details in detail, nor do they limit the invention to only the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of this specification, so that technicians in the relevant technical field can well understand and use this specification. This specification is only limited by the claims and their full scope and equivalents.

Claims

1. A strategy updating method, comprising: Receive a policy update request submitted for a push policy of a target business; Determining an initial push parameter, a push parameter step size, and a plurality of push iteration cycles according to the policy update request; Calculate the push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step size, and determine multiple intermediate push parameters according to the push population information, wherein the push population information is the number of target populations calculated using the initial push parameter in each push iteration cycle, and the intermediate push parameter is the push parameter with the largest number of targeted people in each push iteration cycle; Selecting a target push parameter from the plurality of intermediate push parameters according to a preset convergence function; The push policy is updated based on the target push parameter, and the service user set associated with the target service is determined using the updated push policy.

2. According to the method of claim 1, the receiving a policy update request submitted for a push policy of a target service comprises: Determine a set of subscribed users and a set of unsubscribed users corresponding to the target service, and establish a service push relationship corresponding to the target service according to the set of subscribed users and the set of unsubscribed users; Determining the initial push parameter associated with the target service according to the service push relationship; When an update instruction submitted for the push policy is received, the policy update request corresponding to the push policy is generated according to the policy update parameters carried in the update instruction and the initial push parameters.

3. The method according to claim 1, determining the initial push parameter, the push parameter step size and the multiple push iteration cycles according to the policy update request, comprising: Determine a policy update rule according to the policy update request, and read a push time parameter and a push number parameter according to the policy update rule to form the initial push parameter; The policy update request is parsed to obtain the push parameter step size and the multiple push iteration cycles.

4. The method according to claim 1, wherein the step of calculating the push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step size, and determining a plurality of intermediate push parameters according to the push population information comprises: Using the initial push parameter as the reference push parameter corresponding to the j-th push iteration period; Calculate the push population information corresponding to the j-th push iteration cycle according to the benchmark push parameters; Determine whether the push group information corresponding to the jth push iteration cycle meets the cycle end condition; If not, determine the benchmark push parameter corresponding to the j+1 push iteration period according to the push parameter step size and the benchmark push parameter corresponding to the j-th push iteration period, j is incremented by 1, and execute the step of calculating the push population information corresponding to the j-th push iteration period according to the benchmark push parameter; If so, execute the step of determining multiple intermediate push parameters according to the push group information.

5. The method according to claim 4, further comprising: Determining a reference push parameter step length corresponding to an i-th step length period in the push parameter step length; In the reference push parameter step corresponding to the i-th step period, the initial push parameter is used as the reference push parameter corresponding to the j-th push iteration period; Accordingly, before the step of determining a plurality of intermediate push parameters according to the push group information is performed, the method further includes: Determining whether there are remaining push parameter steps in the push parameter step; If yes, determine the reference push parameter step corresponding to the i+1th step period in the remaining push parameter step, i is incremented by 1, and execute the step of using the initial push parameter as the reference push parameter corresponding to the jth push iteration period in the reference push parameter step corresponding to the i-th step period; If not, execute the step of determining multiple intermediate push parameters according to the push group information.

6. The method according to claim 5, wherein the step of determining a plurality of intermediate push parameters according to the push group information comprises: According to the push group information corresponding to each push iteration cycle, filter the target push group information corresponding to each step cycle; Determine the intermediate push parameters corresponding to each step cycle according to the target push population information corresponding to each step cycle; The plurality of intermediate push parameters are obtained by integrating the intermediate push parameters corresponding to each step period.

7. The method according to claim 1, wherein the step of selecting a target push parameter from the plurality of intermediate push parameters according to a preset convergence function comprises: Determine an initial service user set corresponding to each intermediate push parameter according to the preset convergence function; Calculating an unsubscription rate of the initial service user set according to the number of subscribed users and the number of unsubscribed users included in the initial service user set; A target service user set is determined from the initial service user set corresponding to each intermediate push parameter according to the unsubscription rate, and the intermediate push parameter corresponding to the target service user set is used as the target push parameter.

8. The method according to claim 1, wherein the updating of the push policy based on the target push parameter and determining the set of service users associated with the target service using the updated push policy comprises: Determining service push parameters in the push strategy; updating the service push parameters according to the target push parameters, and determining the target push strategy according to the update result; The target access time and the target access number are determined according to the target push strategy, and the service user set associated with the target service is determined according to the target access time and the target access number.

9. The method according to any one of claims 1 to 8, after the step of updating the push policy based on the target push parameter and determining the set of service users associated with the target service using the updated push policy is executed, further comprising: Determine the push information corresponding to the target service; The push information is sent to the service users included in the service user set according to the updated push policy.

10. A strategy updating device, comprising: A receiving module, configured to receive a policy update request submitted for a push policy of a target service; A determination module, configured to determine an initial push parameter, a push parameter step size, and a plurality of push iteration cycles according to the policy update request; a calculation module configured to calculate the push population information corresponding to each push iteration cycle according to the initial push parameter and the push parameter step size, and determine a plurality of intermediate push parameters according to the push population information, wherein the push population information is the number of target populations calculated using the initial push parameter in each push iteration cycle, and the intermediate push parameter is the push parameter with the largest number of targeted populations in each push iteration cycle; A screening module, configured to screen a target push parameter from the plurality of intermediate push parameters according to a preset convergence function; The updating module is configured to update the push policy based on the target push parameter, and determine the service user set associated with the target service by using the updated push policy.

11. A computing device comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method described in any one of claims 1 to 9 are implemented.

12. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 9.

Citation Information

Patent Citations

  • Resource recommendation and parameter determination method and device, equipment and medium

    CN112163159A

  • Analytic system for machine learning prediction model selection

    US20190258904A1