Strategy determination method and device, equipment, medium and program product

Abstract

The invention discloses a strategy determination method and device, equipment, a medium and a program product, and relates to the technical field of computers. The method comprises the following steps: acquiring object data information corresponding to a plurality of objects respectively; performing feature analysis on the object data information corresponding to the plurality of objects to obtain activity prediction results corresponding to the plurality of objects; determining a resource allocation strategy corresponding to the plurality of objects; based on the activity prediction results corresponding to the plurality of objects and a resource allocation strategy, determining efficiency saturation; and under the condition that the efficiency saturation meets a preset saturation condition, determining the resource allocation strategy as a target allocation strategy. According to the embodiment of the invention, the efficiency saturation for measuring the predicted income is determined based on the activity prediction result and the resource allocation strategy, and the target allocation strategy is determined under the constraint of the preset saturation condition, so that the activity resources can be concentrated on an object with higher expected income, the maximum utilization of the activity resources is ensured, and the waste of the activity resources is avoided.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, and in particular to a policy determination method, apparatus, device, medium, and program product. Background Art

[0002] Some game activities will be carried out in the game application, such as: New Year welfare activities held during the Spring Festival, character welfare activities held when new characters are launched, etc.

[0003] In related technologies, before a game activity is launched, game operators need to formulate a reasonable resource allocation strategy (i.e., resources allocated to each player) based on the total resources of the activity. Usually, game operators will refer to the resource allocation strategies of historical activities with better returns in the game application to determine the resource allocation strategy for the activity to be launched.

[0004] However, since historical activities with better returns may also result in waste of allocated resources, if the resource allocation strategy for future activities is guided by the input costs of historical activities with better returns, it is inevitable that future activities will also result in waste of allocated resources. Summary of the Invention

[0005] The embodiments of the present application provide a policy determination method, apparatus, device, medium, and program product. The technical solutions are as follows:

[0006] In one aspect, a policy determination method is provided, the method comprising:

[0007] Obtaining object data information corresponding to a plurality of objects, wherein the object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period;

[0008] Performing feature analysis on object data information corresponding to each of the plurality of objects to obtain activity prediction results corresponding to each of the plurality of objects, the activity prediction results being used to characterize the predicted participation of the objects in a target activity, the target activity comprising an activity to be carried out on the first interactive platform;

[0009] Determining resource allocation strategies corresponding to the multiple objects, the resource allocation strategies being used to indicate allocation of preset resources delivered by the first interactive platform among the multiple objects;

[0010] Determining, based on the activity prediction results corresponding to the plurality of objects and the resource allocation strategy, a performance saturation, wherein the performance saturation is used to represent the predicted benefit corresponding to the target activity when the resource allocation strategy is executed;

[0011] In a case where the performance saturation meets a preset saturation condition, the resource allocation strategy is determined as a target allocation strategy, and the target allocation strategy is used to allocate the preset resources among the multiple objects.

[0012] In another aspect, a policy determination device is provided, the device comprising:

[0013] An acquisition module, configured to acquire object data information corresponding to a plurality of objects, wherein the object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period;

[0014] an analysis module, configured to perform feature analysis on the object data information corresponding to the plurality of objects, to obtain activity prediction results corresponding to the plurality of objects, the activity prediction results being used to characterize the predicted participation of the objects in a target activity, the target activity comprising an activity to be carried out on the first interactive platform;

[0015] a determination module, configured to determine resource allocation strategies corresponding to the plurality of objects, wherein the resource allocation strategies are configured to indicate allocation of preset resources delivered by the first interactive platform among the plurality of objects;

[0016] The determining module is further configured to determine a performance saturation based on the activity prediction results corresponding to the plurality of objects and the resource allocation strategy, wherein the performance saturation is used to represent the predicted benefit corresponding to the target activity when the resource allocation strategy is executed;

[0017] The determining module is configured to determine the resource allocation strategy as a target allocation strategy when the performance saturation satisfies a preset saturation condition, wherein the target allocation strategy is used to allocate the preset resources among the multiple objects.

[0018] On the other hand, a computer device is provided, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement any of the above-mentioned policy determination methods.

[0019] On the other hand, a computer-readable storage medium is provided, wherein the storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, the at least one program, the code set or instruction set is loaded and executed by a processor to implement any of the above-mentioned policy determination methods.

[0020] In another aspect, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform any of the aforementioned policy determination methods.

[0021] The beneficial effects of the technical solutions provided in the embodiments of the present application include at least:

[0022] By analyzing the characteristics of the object data corresponding to multiple objects, the target activity participation status of each of the multiple objects is determined. Then, based on the participation status of the multiple objects and the determined resource allocation strategy, the performance saturation is determined. If the performance saturation meets the preset saturation conditions, the resource allocation strategy is determined as the target allocation strategy, and the preset resources are allocated among the multiple objects according to the target allocation strategy. By determining the performance saturation used to measure the predicted benefits based on the activity prediction results and the resource allocation strategy, and determining the target allocation strategy under the preset saturation conditions, it is possible to concentrate activity resources on objects with higher expected benefits, ensuring maximum utilization of activity resources and avoiding waste of activity resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0024] Figure 1 is a schematic diagram of an implementation environment provided by an exemplary embodiment of the present application;

[0025] Figure 2 is a flow chart of a policy determination method provided by an exemplary embodiment of the present application;

[0026] Figure 3 is a flow chart of a policy determination method provided by another exemplary embodiment of the present application;

[0027] Figure 4 is a flowchart of a policy determination method provided by yet another exemplary embodiment of the present application;

[0028] Figure 5 is a schematic diagram of a function image provided by an exemplary embodiment of the present application;

[0029] Figure 6is a flowchart of a policy determination method provided by another exemplary embodiment of the present application;

[0030] Figure 7 is a structural block diagram of a policy determination device provided by an exemplary embodiment of the present application;

[0031] Figure 8 is a structural block diagram of a policy determination device provided by another exemplary embodiment of the present application;

[0032] Figure 9 It is a structural block diagram of a computer device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0033] To make the objectives, technical solutions, and advantages of this application more clear, the following will further describe the embodiments of this application in detail with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0034] In this application, the terms "first", "second", etc. are used to distinguish identical or similar items with substantially the same effects and functions. It should be understood that there is no logical or temporal dependency between "first" and "second", nor is there any limitation on the quantity and execution order.

[0035] In related technologies, before launching a game activity, game operators need to formulate a reasonable resource allocation strategy (i.e., the resources allocated to each player) based on the total resources of the activity. Typically, game operators refer to the resource allocation strategies of historical activities with high returns within the game application to determine the resource allocation strategy for the upcoming activity. However, since historical activities with high returns can also waste allocated resources, using the investment costs of historical activities with high returns to guide the resource allocation strategy for future activities will inevitably lead to resource waste in future activities.

[0036] An embodiment of the present application provides a strategy determination method, which determines the performance saturation used to measure the predicted benefits based on the activity prediction results and the resource allocation strategy, and determines the target allocation strategy under the constraints of preset saturation conditions. This can concentrate activity resources on objects with higher expected benefits, ensure that activity resources are maximized, and avoid wasting activity resources.

[0037] The strategy determination method provided in the embodiments of the present application can be applied to activity resource allocation scenarios, advertising promotion scenarios, object recall scenarios, etc., and the embodiments of the present application are not limited to this.

[0038] Next, the implementation environment involved in the embodiments of the present application is described.

[0039] The policy determination method provided in the embodiment of the present application can be implemented by the terminal alone, or by the server, or by the terminal and the server through data interaction, which is not limited in the embodiment of the present application. Optionally, the policy determination method implemented by the terminal and the server through data interaction is used as an example for description.

[0040] For illustration, please refer to Figure 1 , the implementation environment involves a terminal 110 and a server 120, and the terminal 110 and the server 120 are connected via a communication network 130. Optionally, the communication network 130 can be a wired network or a wireless network, which is not limited in this embodiment of the present application.

[0041] There may be one or more terminals 110. A first interactive platform may be installed in the terminal 110. The first interactive platform may be any of the following types of applications, such as game applications, video applications, lifestyle applications, social applications, etc. Optionally, the application may be an application that requires downloading and installation, or an application that can be used instantly, which is not limited in this embodiment of the present application.

[0042] The server 120 is used to provide background services for the installation and operation of the first interactive platform in the terminal 110. For example, the server 120 may be a background server of the aforementioned game application.

[0043] Illustratively, the server 120 obtains object data information corresponding to n objects from the terminal 110, where n is an integer greater than 1, and the object data information includes historical activity data corresponding to the first interactive platform. The server 120 performs feature analysis on the object data information corresponding to the n objects to obtain activity prediction results corresponding to the n objects, such as the probability of the objects participating in the target activity to be carried out on the first interactive platform. The server 120 then determines resource allocation strategies corresponding to the multiple objects and determines performance saturation based on the activity prediction results and resource allocation strategies corresponding to the multiple objects. If the performance saturation meets the preset saturation condition, the server 120 determines the resource allocation strategy as the target allocation strategy, which is the resource allocation strategy ultimately formulated for the target activity. Optionally, if the performance saturation meets the preset saturation condition, the server 120 re-formulates the resource allocation strategy and performs performance saturation calculation and preset saturation condition judgment.

[0044] Optionally, the server 120 allocates preset resources among multiple objects according to the target allocation policy.

[0045] It is worth noting that the above-mentioned terminal 110 includes but is not limited to mobile terminals such as mobile phones, tablet computers, portable laptops, intelligent voice interaction devices, smart home appliances, and vehicle-mounted terminals, and can also be implemented as desktop computers, etc.; the server 120 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, and can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network) and big data and artificial intelligence platforms.

[0046] Cloud technology refers to a managed technology that unifies hardware, software, and network resources within a wide area network (WAN) or local area network (LAN) to enable data computing, storage, processing, and sharing. Cloud technology is a general term for network, information technology, integration technology, management platform technology, and application technology, all based on the cloud computing business model. It can form a resource pool for on-demand, flexible, and convenient use. Cloud computing technology will become a crucial support. Backend services in technical network systems, such as those for video websites, image websites, and more portals, require significant computing and storage resources. With the rapid development and application of the internet industry, every item will likely have its own unique identifier, requiring transmission to backend systems for logical processing. Data of varying levels will be processed separately, requiring robust system support for all types of industry data, which can only be achieved through cloud computing. Alternatively, server 120 can also be implemented as a node in a blockchain system.

[0047] It should be noted that before collecting the user's relevant data and during the process of collecting the user's relevant data, this application can display a prompt interface, pop-up window or output voice prompt information. The prompt interface, pop-up window or voice prompt information is used to remind the user that its relevant data is currently being collected, so that this application only starts to execute the relevant steps of obtaining the user's relevant data after obtaining the user's confirmation operation on the prompt interface or pop-up window. Otherwise (that is, when the user's confirmation operation on the prompt interface or pop-up window is not obtained), the relevant steps of obtaining the user's relevant data are terminated, that is, the user's relevant data is not obtained. In other words, all user data collected by this application are collected with the user's consent and authorization, and the collection, use and processing of relevant user data need to comply with relevant laws, regulations and standards.

[0048] Next, the strategy determination method provided by this application is described.

[0049] Combined with the above introduction, Figure 2 This is a flow chart of a strategy determination method provided by an embodiment of the present application, in which the method is applied to Figure 1Taking the server shown as an example, the method is as follows: steps 210 to 250.

[0050] Step 210: Obtain object data information corresponding to a plurality of objects.

[0051] Optionally, the multiple objects refer to objects in the first interactive platform, for example, the multiple objects refer to multiple registered accounts in the first interactive platform; or, the multiple objects refer to objects in the first interactive platform that have permission to participate in the target activity.

[0052] Optionally, the first interactive platform includes game applications, video applications, life applications, social applications, etc. The following mainly describes the case where the first interactive platform is implemented as a game application.

[0053] Illustratively, the object data information refers to information generated by the object using the first interactive platform. Optionally, the object data information includes at least one of the following information:

[0054] 1. Historical activity data.

[0055] Historical activity data is used to indicate the subject's participation in activities on the first interactive platform within a historical time period: the number of activities participated in in the past three months, the length of stay in the activities, the number of activity products collected, the number of product clicks, the number of product purchases, the number of gift packs collected at each level, the number of coupons collected at each level, the number of tasks completed, the number of gifts given away, the number of return logins to the game after the activity, the number of return purchases, etc.

[0056] The end time of the above-mentioned historical time period can be the current time or any historical time before the current time. The length of the historical time period includes 1 year, half a year, 3 months, etc., which is not limited in the embodiments of this application.

[0057] 2. Historical login data.

[0058] Historical login data indicates the subject's login status on the First Interactive Platform during a historical period. Examples include: the number of game login days per month over the past three months, the monthly month-over-month growth rate of the number of login days, the monthly online game time spent per month over the past three months, the monthly month-over-month growth rate of online time spent per month, the number of game logins per month over the past three months, the monthly month-over-month growth rate of game logins, the increase in the average number of weekend logins per month compared to the average number of weekday logins per month over the past three months, and the percentage of login time spent per time period per month over the past three months.

[0059] 3. Object attribute data.

[0060] The object attribute data is used to indicate the basic attributes of the object in the first interactive platform, such as the object's level in the game, game nickname, login terminal type, current activity level, registration time, etc.

[0061] 4. Object social data.

[0062] The subject's social data is used to describe the subject's social interactions on the first interactive platform, such as the subject's in-game friendships, guild / team affiliation, etc.

[0063] 5. Object feedback data.

[0064] The subject feedback data is used to indicate the data feedback submitted by the subject in the first interactive platform, such as the subject's evaluation, comments, feedback, and complaints in the game.

[0065] It should be noted that the above examples of object data information are only illustrative and are not limited to these in the embodiments of the present application.

[0066] In some embodiments, the object data information corresponding to the above-mentioned multiple objects can be obtained from the terminals logged in by the multiple objects respectively.

[0067] Optionally, the terminals that multiple objects log in to correspond to a management terminal, where the management terminal refers to a terminal that provides a target application, including but not limited to: an activity management platform, a cost control platform, and other platforms with activity resource allocation and policy-making functions. For example, a game operator logs into the activity management platform through the management terminal. On the activity management platform, the operator can determine the target objects of the target activity (i.e., determine multiple objects with participation permissions) and specify the data information of the objects participating in the activity prediction results in the activity management platform.

[0068] Step 220 : performing feature analysis on the object data information corresponding to the plurality of objects to obtain activity prediction results corresponding to the plurality of objects.

[0069] The activity prediction result is used to characterize the predicted participation of the subject in the target activity, and the target activity includes activities to be carried out on the first interactive platform.

[0070] Illustratively, the target activity refers to a specific activity to be carried out by the first interactive platform, that is, the target activity specifically refers to a certain activity; or, the target activity generally refers to an activity to be held in the future by the first interactive platform, which is not limited in this embodiment of the present application.

[0071] Optionally, the activity prediction result includes at least one of the following results:

[0072] 1. Activity participation probability, which is used to represent the predicted probability of an object participating in the target activity.

[0073] 2. Activity prediction results include activity consumption probability, which is used to represent the predicted probability of the subject participating in the target activity and consuming.

[0074] 3. Predicted activity,Activity activity is used to characterize the predicted activity of the,object in the target activity.

[0075] It should be noted that the above examples of activity prediction results are only for illustrative purposes and are not limited to these examples in the present application.

[0076] Optionally, feature analysis is performed on object data information corresponding to the plurality of objects using an activity prediction model to obtain activity prediction results corresponding to the plurality of objects.

[0077] Illustratively, the activity prediction model can be implemented as at least one of a logistic regression model, a convolutional neural network (CNN), or a recurrent neural network (RNN), etc., and this embodiment of the present application does not limit this.

[0078] Optionally, the activity prediction model is used to perform feature analysis on the object data information of a single object to obtain the activity prediction result of the single object; or, the activity prediction model is used to perform feature analysis on the object data information corresponding to multiple objects as a whole to obtain the activity prediction results corresponding to multiple objects.

[0079] Optionally, feature engineering processing is performed on object data information corresponding to the multiple objects to obtain object feature representations corresponding to the multiple objects; and the object feature representations corresponding to the multiple objects are predicted through an activity prediction model to obtain activity prediction results corresponding to the multiple objects.

[0080] Illustratively, the object feature representation includes at least one of an activity feature representation corresponding to historical activity data, a login feature representation corresponding to historical login data, an attribute feature representation corresponding to object attribute data, a social feature representation corresponding to object social data, and a feedback feature representation corresponding to object feedback data, etc., and the embodiments of the present application do not limit this.

[0081] Optionally, the above-mentioned feature engineering processing includes at least one of feature preprocessing, feature extraction, feature selection, feature construction, etc., which is not limited in the embodiment of the present application.

[0082] Step 230: Determine resource allocation strategies corresponding to the multiple objects.

[0083] In principle, pre-set resources refer to the resources invested in carrying out target activities.

[0084] For example, the preset resources refer to the resource investment amount corresponding to the target activity. This resource investment amount can be used to deploy various types of resources. Taking game applications as an example, the resources deployed include virtual gold coins, virtual characters, virtual props, virtual decorations, coupons, memberships, etc. that can be used on the first interactive platform; the resources deployed also include promotional resources, such as coupons and memberships in other applications; the resources deployed also include physical resources, such as physical peripherals of game characters, etc.

[0085] The resource allocation strategy is used to indicate the allocation of preset resources launched by the first interactive platform among multiple objects. For example, the resource investment quota of the target activity launched by the first interactive platform is 10,000, and the quota allocated to object A among the multiple objects is 100.

[0086] Step 240 : determining performance saturation based on the activity prediction results and resource allocation strategies corresponding to the plurality of objects.

[0087] Among them, performance saturation is used to characterize the predicted benefits corresponding to the target activity when the resource allocation strategy is executed.

[0088] Optionally, the predicted revenue includes activity participation, wherein activity participation refers to the number of objects that may participate in the target activity, and the higher the efficacy saturation, the more objects participate in the activity.

[0089] Optionally, the predicted revenue also includes activity consumption, where activity consumption refers to the consumption of participants in the target activity. The higher the performance saturation, the more participants participate in the target activity and make consumption.

[0090] Step 250 : When the performance saturation meets the preset saturation condition, the resource allocation strategy is determined as the target allocation strategy.

[0091] In some embodiments, when the performance saturation is greater than or equal to a preset saturation, the resource allocation strategy is determined as the target allocation strategy.

[0092] Schematically, a preset saturation is set, and the performance saturation calculation is performed for the currently determined resource allocation strategy. If the performance saturation of the resource allocation strategy is less than the preset saturation, the next resource allocation strategy is determined, and the performance saturation calculation is performed for the next resource allocation strategy, and the obtained performance saturation is compared with the preset saturation, until the performance saturation of the determined resource allocation strategy is greater than or equal to the preset saturation, and the determined resource allocation strategy is used as the target allocation strategy.

[0093] In the above embodiment, the preset saturations of the resource allocation strategies are calculated in turn. If the preset saturation of a certain resource allocation strategy is calculated to be greater than the preset saturation, the resource allocation strategy is determined as the target allocation strategy. That is, the saturation calculation can be stopped after a solution that meets the conditions is found, thereby saving computing resources.

[0094] Alternatively, when the performance saturation reaches the target saturation, the resource allocation strategy is determined as the target allocation strategy.

[0095] The target saturation refers to the maximum performance saturation that can be achieved based on the activity prediction results corresponding to multiple objects under the condition of fixed preset resources.

[0096] Schematically, all possible resource allocation strategies are determined based on the preset number of resources and the number of objects. For example, if the current preset number of resources is 200 and the number of objects is 100, the number of resources that each object can obtain can be any integer between 0 and 200. Therefore, the number of all possible resource allocation resources is 201 100 The efficiency saturation is calculated for all possible resource allocation strategies, and the efficiency saturation corresponding to each of the possible resource allocation strategies is obtained. The resource allocation strategy corresponding to the maximum efficiency saturation is taken as the target allocation strategy.

[0097] In the above embodiment, the efficiency saturation of each resource allocation strategy is calculated, and the resource allocation strategy corresponding to the maximum efficiency saturation is determined as the target allocation strategy, thereby ensuring the maximization of the predicted benefit.

[0098] The target allocation strategy is used to allocate preset resources among multiple objects.

[0099] Optionally, when the target activity is enabled, preset resources are allocated among the multiple objects according to the target allocation strategy.

[0100] In an illustrative manner, when a target activity is carried out, preset resources can be allocated among multiple objects according to the target allocation strategy. For example, if the allocation quota of object 1 in the preset resource quantity target allocation strategy is 100, then resources will be allocated to object 1 according to the resource quantity of 100, for example, providing object 1 with a coupon with a quota of 100, etc.

[0101] In some embodiments, when a target activity is started and a target object among multiple objects logs into the first interactive platform, target resources are delivered to the target object according to a resource allocation result of the target object in the target allocation strategy.

[0102] The target resources include at least one of activity resources, physical resources, and promotional resources.

[0103] Activity resources include resources provided to participants for use in the target activity, such as limited-edition virtual items. Promotional resources include resources provided to participants for use on other interactive platforms, such as game skins for other games associated with the game currently hosting the activity.

[0104] Illustratively, after the target activity is started, when the target object logs in to the first interactive platform, resources are delivered to the target object according to the target object's allocation quota in the target allocation strategy. The delivered resources can be pre-set by the operator; or, the delivered resources are selected by the object in the first interactive platform (for example, providing the object with character coupons and prop coupons, and the object can choose one from them). This embodiment of the present application is not limited to this.

[0105] In the above embodiment, resource delivery is performed when the object logs in after the target activity is started. Compared with delivering resources to each object immediately after the target activity is started, it can ensure that the delivered resources can accurately reach the target object and improve the efficiency of resource use.

[0106] Optionally, when the target activity is started and the target object has not logged into the first interactive platform within a preset time period, the target resources are delivered to the designated object according to the resource allocation result of the target object in the target allocation strategy.

[0107] The preset time period is the time period starting from the start time of the target activity, and the duration of the preset time period is less than or equal to the duration of the target activity. For example, if the duration of the target activity is 10 days, the preset time period is within 9 days after the target activity starts.

[0108] Optionally, the designated object refers to any object currently logged into the first interactive platform, or the designated object refers to an object whose activity in the target activity is greater than or equal to a preset activity level. That is, if the target object does not log into the first interactive platform for a period of time after the target activity is started, the resources originally intended for the target object may be allocated to other objects.

[0109] In the above embodiment, after the target activity is started, if the target object has not logged into the first interactive platform for a long time, the resources originally intended for the target object are delivered to other objects that have logged into the platform, thereby avoiding waste of resources and improving resource utilization.

[0110] In summary, the embodiment of the present application provides a strategy determination method, which obtains the target activity participation status corresponding to the multiple objects by performing feature analysis on the object data information corresponding to the multiple objects respectively; then, based on the participation status corresponding to the multiple objects and the determined resource allocation strategy, the performance saturation is determined, and when the performance saturation meets the preset saturation condition, the resource allocation strategy is determined as the target allocation strategy, so that the preset resources are allocated among the multiple objects according to the target allocation strategy. Based on the activity prediction results and the resource allocation strategy, the performance saturation used to measure the predicted benefits is determined, and the target allocation strategy is determined under the constraints of the preset saturation condition, so that the activity resources can be concentrated on the objects with higher expected benefits, ensuring that the activity resources are maximized and avoiding the waste of activity resources.

[0111] In some embodiments, a logistic regression model is used to perform feature analysis on the object data information corresponding to the multiple objects to obtain the activity prediction results corresponding to the multiple objects. Figure 3 As shown above Figure 2 The illustrated embodiment may also be implemented as the following steps 310 to 350 .

[0112] Step 310: Obtain object data information corresponding to a plurality of objects.

[0113] The object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period.

[0114] Optionally, candidate object data information corresponding to the plurality of objects is obtained; and data cleaning processing is performed on the candidate object data information corresponding to the plurality of objects to obtain object data information corresponding to the plurality of objects.

[0115] Illustratively, after obtaining the candidate object data information of the object, data cleaning is performed on the information, thereby improving the quality of the object data information finally obtained, which is beneficial to the subsequent activity participation prediction process.

[0116] Optionally, the data cleaning process includes at least one of data deduplication, outlier detection and processing, etc., which is not limited in the embodiment of the present application.

[0117] Step 320 : Perform feature analysis on the object data information corresponding to the multiple objects using a logistic regression model to obtain activity prediction results corresponding to the multiple objects.

[0118] The activity prediction result is used to characterize the predicted participation of the subject in the target activity, and the target activity includes activities to be carried out on the first interactive platform.

[0119] Schematically, a logistic regression model is a machine learning method used to solve binary classification problems. The logistic regression model can map continuous input values ​​between 0 and 1 to represent the probability of an event (e.g., a subject participating in a target activity).

[0120] The basic principles of the logistic regression model are explained below:

[0121] Logistic regression models are faced with two choices, such as choosing y = 1 (participation) or y = 0 (non-participation) when faced with a target activity. This is the explained variable y. This choice is influenced by a variety of factors, which are the explanatory variables. These explanatory variables are included in the vector x.

[0122] For the explained variable y i , y i Treated as a random variable Y i The realized value, Y i The probability of taking 1 is s, Y i The probability of taking 0 is 1-s. Random variable Y i Obey the (0-1) distribution with parameter s, Y i The distribution law is as follows:

[0123] Formula 1:

[0124] Obviously, if y i =1, then Y i The probability is s; if y i =0, then Y i The probability is 1-s; i refers to the i-th object.

[0125] From the above, we can see that the problem of explaining the choice y under the influence of multiple factors x becomes the problem of explaining the probability s under the influence of multiple factors x. If the relationship between the two is set as a linear function as shown in the following formula 2:

[0126] Formula 2: s i =x i ×β

[0127] Among them, β is the preset parameter, which means x i and a i The correlation between them.

[0128] This model is often called a linear probability model. To calculate the probability s, it is necessary to transform the probability s to remove the constraints on its range of values, and then set the transformed value as a linear function of the explanatory variable x. This process can be divided into two steps.

[0129] The first step is to define the odds ratio (the ratio of the probability of an event occurring to the probability of the event not occurring) based on the probability s. The formula for calculating the odds ratio is shown in Formula 3 below:

[0130] Formula 3:

[0131] That is Y i =1 probability s i With Y i =0 probability 1-s i ratio.

[0132] Obviously, the odds ratio can take any non-negative value, thus eliminating the upper limit constraint on the value range.

[0133] The second step is to take the logarithm to calculate the logit or log-odds (the logarithm of the odds ratio is called logit or log-odds). The formula for calculating logit or log-odds is shown in Formula 4 below:

[0134] Formula 4:

[0135] This will also eliminate the lower limit constraint. Through the above transformation, logitΩ i Map the original value range from probability s (0, 1) to the entire real number range. At this point, the derivation of the logistic regression model is complete.

[0136] Optionally, feature engineering processing is performed on object data information corresponding to the multiple objects to obtain object feature representations corresponding to the multiple objects; and the object feature representations corresponding to the multiple objects are predicted through a logistic regression model to obtain activity prediction results corresponding to the multiple objects.

[0137] Among them, feature engineering processing includes at least one of feature preprocessing, feature extraction, feature selection, feature construction, etc., which is not limited in the embodiments of the present application.

[0138] Feature preprocessing includes at least one of missing value repair, discretization, and normalization. The following describes missing value repair, discretization, and normalization, respectively.

[0139] Missing value repair refers to the processing of missing values ​​in data. Missing value repair includes at least one of the following: deletion (when the proportion of missing values ​​in a variable is too large, you can consider directly deleting the variable), filling (using a statistical value such as the mean, median, mode, etc. or a specific value to replace the missing value), prediction (using other variables to establish a prediction model and use the predicted value to fill the missing value), and interpolation (using interpolation methods such as linear interpolation and polynomial interpolation to estimate the missing value).

[0140] For example, in this embodiment, considering data integrity, if a variable of an object is missing, the missing value is filled with the average value of the variable. For example, if the grade of object 1 is missing, the average grade of multiple objects is used as the grade of object 1.

[0141] Discretization is the process of converting continuous variables into categorical variables. Discretization can include at least one of equal-width processing (dividing the variable based on its range so that each interval has equal width), equal-frequency processing (dividing the variable based on the frequency of occurrence so that each interval has equal number of samples), and clustering (using the K-means algorithm to divide the variable into multiple intervals based on a specified number of intervals).

[0142] Schematically, the purpose of discretization is to increase the nonlinearity of the model and improve the generalization ability of the model. In the embodiment of the present application, features such as level are discretized using equal-width processing. The specific operation is to group the features according to the same principle of each grouping interval. For example, the value range of the object game level is 1 to 80, and the equal-width processing takes 10 as the interval interval, and the game level is divided into (1, 10), (11, 20), (21, 30), (31, 40), (41, 50), (51, 60), (61, 70), and (71, 80) in sequence.

[0143] Normalization is the process of converting dimensional expressions into dimensionless ones through linear transformation. The purpose of normalization is to eliminate the influence of dimensions between metrics and make them comparable. Normalized data can also speed up the gradient descent method in finding the optimal solution.

[0144] Schematically, the normalized calculation formula is shown in Formula 5 below:

[0145] Formula 5:

[0146] Among them, X represents the original feature, X min represents the minimum value of the original feature, X max represents the maximum value of the original feature, and x′ is the feature after normalization.

[0147] After the above feature engineering processing, the object feature representations corresponding to multiple objects can be obtained. By inputting the object feature representations into the calculation formula corresponding to the logistic regression model, the activity prediction results corresponding to the multiple objects can be obtained.

[0148] Schematically, the calculation formula of the logistic regression model is shown in Formula 6 below:

[0149] Formula 6:

[0150] Where x represents the object feature representation, w is the weight parameter, and e is a natural constant. P is used to measure the object's participation in the target activity, for example, the probability of the object participating in the target activity.

[0151] Where w is the weight of the model learning. The value of w can be obtained by training the sample logistic regression model by setting the initial value. Schematically, the training scheme of the sample logistic regression model is described as follows:

[0152] A sample data set is obtained, where the sample data set includes sample object data information of the object, and the sample object data information includes: sample historical activity data, sample historical login data, sample object attribute data, etc., wherein the sample object data information corresponds to a reference activity participation result; feature engineering processing is performed on the sample object data information to obtain a feature representation of the sample object; the sample object feature representation is predicted through a logistic regression model to obtain a predicted activity participation result of the object; and a sample logistic regression model is trained based on the difference between the reference activity participation result and the predicted activity participation result to obtain a logistic regression model.

[0153] For illustration, the reference activity participation result is realized as the reference activity participation probability, and the predicted activity participation result is realized as the predicted activity participation probability. The reference activity participation probability refers to the actual probability that the sample object participates in the sample activity (an activity that has been carried out), which is generally 0 or 1. After the predicted activity participation probability is predicted, the model loss is calculated based on the difference between the reference activity participation probability and the predicted activity participation probability. The parameters of the sample logistic regression model are updated through the loss to obtain a logistic regression model.

[0154] The number of subjects mentioned above is a single one, meaning that the sample logistic regression model is trained using the sample subject data corresponding to a single subject. This results in a dedicated logistic regression model, which is used only to predict the activity participation of a single subject. The number of sample subject data corresponding to a single subject can be one or more.

[0155] Alternatively, the number of objects mentioned above is multiple, that is, the sample logistic regression model is trained using the sample object data information corresponding to each of the multiple objects. A logistic regression model shared by the multiple objects is obtained, and this logistic regression model can be used to predict the activity participation of the multiple objects. The number of sample object data information corresponding to each of the multiple objects can be one or more.

[0156] In some embodiments, when the target activity is a specific activity, an activity feature representation corresponding to the target activity is obtained. For example, activity information of the target activity is obtained, including the activity type, activity description, activity promotional images, activity rules, activity process, and activity rewards of the target activity; and feature extraction is performed on the activity information to obtain the aforementioned activity feature representation.

[0157] Optionally, the object feature representations and activity feature representations respectively corresponding to the multiple objects are predicted using a logistic regression model to obtain activity prediction results respectively corresponding to the multiple objects.

[0158] Schematically, when predicting activity prediction results, in addition to considering the subject's object characteristics, we also consider the activity feature representation of the upcoming activity. Activity feature representations typically represent various attributes and information about the activity, which are directly related to the user's participation intention and behavior patterns. By incorporating these activity feature representations into the logistic regression model, we can more accurately capture the subject's preference and responsiveness to the target activity, thereby increasing the probability of the predicted activity prediction results.

[0159] Step 330: Determine resource allocation strategies corresponding to the multiple objects.

[0160] The resource allocation strategy is used to indicate the allocation of preset resources delivered by the first interactive platform among multiple objects.

[0161] Step 340 : Determine the performance saturation based on the activity prediction results and resource allocation strategies corresponding to the multiple objects.

[0162] Among them, performance saturation is used to characterize the predicted benefits corresponding to the target activity when the resource allocation strategy is executed.

[0163] Step 350 : When the performance saturation meets the preset saturation condition, the resource allocation strategy is determined as the target allocation strategy.

[0164] The target allocation strategy is used to allocate preset resources among multiple objects.

[0165] In summary, the strategy determination method provided in the embodiments of this application utilizes a logistic regression model to predict the object feature representations corresponding to multiple objects. This model boasts rapid training and prediction speeds, improving the efficiency of activity prediction results and, consequently, the effectiveness of target allocation strategy formulation. Furthermore, prior to prediction, the object data information corresponding to the multiple objects is preprocessed using multiple feature preprocessing methods, improving the quality of the resulting object feature representations and, consequently, the accuracy of the resulting activity prediction results.

[0166] In some embodiments, the example of using performance saturation to characterize the activity participation of a target activity when executing a resource allocation strategy is described. Figure 4 As shown above Figure 2 or Figure 3 The illustrated embodiment may also be implemented as the following steps 401 to 405 .

[0167] Step 401: Obtain object data information corresponding to a plurality of objects.

[0168] The object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period.

[0169] Step 402 : performing feature analysis on the object data information corresponding to the plurality of objects to obtain activity participation probabilities corresponding to the plurality of objects.

[0170] The activity participation probability is used to represent the predicted probability of the subject participating in the target activity, and the target activity includes activities to be carried out on the first interactive platform.

[0171] Optionally, feature engineering processing is performed on object data information corresponding to the multiple objects to obtain object feature representations corresponding to the multiple objects; and the object feature representations corresponding to the multiple objects are predicted through a logistic regression model to obtain activity participation probabilities corresponding to the multiple objects.

[0172] Schematically, the calculation formula of the logistic regression model for calculating the probability of activity participation is shown in Formula 7 below:

[0173] Formula 7:

[0174] Among them, n is the number of objects, a i and b i is the weight parameter in the logistic regression model (i.e., the parameter learned during the logistic regression model training), x i refers to the feature representation of the i-th object, P i It refers to the probability that the i-th subject participates in the target activity.

[0175] Step 403: Determine the resource allocation quotas corresponding to the multiple objects.

[0176] The resource allocation quota is used to indicate the allocation quota of the preset resources put in by the first interactive platform among the multiple objects. The preset resources include the resource investment quota corresponding to the target activity.

[0177] For example, the sum of the allocations to multiple objects is the resource investment quota. For example, if the current fixed preset resource quantity is 350, and the multiple objects are Object 1, Object 2, and Object 3, then 100 resources can be allocated to Object 1, 50 resources to Object 2, and 200 resources to Object 3. The total of 100, 50, and 200 is the resource allocation quota.

[0178] Step 404 : performing weighted summation on the resource allocation quotas corresponding to the multiple objects based on the activity participation probabilities corresponding to the multiple objects, to obtain performance saturation.

[0179] Among them, efficacy saturation is positively correlated with the number of participants in the target activity.

[0180] Optionally, the calculation formula for the efficacy saturation of a single object is shown in the following formula 8:

[0181] Formula 8: d i =D i ×P i

[0182] Among them, D i refers to the resource allocation quota of the i-th object, d i It refers to the performance saturation corresponding to the i-th object. The performance saturation of n objects can be obtained by adding up the performance saturation of n objects.

[0183] Schematically, the current activity participation probabilities of objects 1, 2, and 3 are 0.5, 0.1, and 0.8, respectively. The current fixed preset number of resources is 350. The resource allocation strategy includes an allocation quota of 100 for object 1, 50 for object 2, and 200 for object 3. The calculated performance saturation is 215.

[0184] It should be noted that the above resource allocation quota can also be implemented as a resource allocation ratio.

[0185] For example, the resource allocation strategy includes an allocation ratio of 2 / 7 for object 1, an allocation quota of 50 for object 2, and an allocation ratio of 1 / 7 for object 3. The calculated efficiency saturation is 4 / 7. The calculated efficiency saturation is 43 / 70.

[0186] In some embodiments, the activity prediction result further includes a predicted activity level, where the activity level is used to characterize the predicted activity level of the object in the target activity.

[0187] Optionally, multiple objects are grouped according to m activity intervals to obtain m groups of objects, where the predicted activity of the i-th group of objects is in the i-th activity interval, m is an integer greater than 1, i≤m and i is a positive integer; the activity weights corresponding to the m groups of objects are determined, and the activity weights are positively correlated with the activity sizes corresponding to the activity intervals; based on the activity participation probabilities corresponding to the multiple objects and the activity weights corresponding to the m groups of objects, the resource allocation amounts or resource allocation ratios corresponding to the multiple objects are weightedly summed to obtain performance saturation.

[0188] For example, assuming that the m activity intervals are [0, 30], [31, 80], and [81, 100], and the predicted activities of objects 1 to 5 are 30, 2, 99, 62, and 15, respectively, after grouping according to the predicted activity, objects 1, 2, and 5 belong to group 1, with the corresponding activity interval being [0, 30]; object 4 belongs to group 2, with the corresponding activity interval being [31, 80]; and object 3 belongs to group 3, with the corresponding activity interval being [81, 100].

[0189] Assign a weight to each activity interval. For example, the weight of [0, 30] is 0.2, the weight of [31, 80] is 0.3, and the weight of [81, 100] is 0.5. Assuming that the current activity participation probabilities of objects 1 to 5 are 0.5, 0.1, 0.8, 0.4, and 0.6, respectively, and the allocation quotas of objects 1 to 5 in the resource allocation strategy are 100, 50, 200, 150, and 60, respectively, the final performance saturation is 0.2 × 0.5 × 100 (object 1) + 0.2 × 0.1 × 50 (object 2) + 0.5 × 0.8 × 200 (object 3) + 0.3 × 0.4 × 150 (object 4) + 0.2 × 0.6 × 60 (object 5) = 116.2.

[0190] In the above embodiment, it is taken into account that there are objects with a high probability of participating in activities but low activity activity; or, objects with a low probability of participating in activities but high activity activity. Therefore, when calculating the performance saturation, the activity activity information is taken into account to more comprehensively evaluate the actual contribution and participation of the object, thereby obtaining a more accurate performance saturation calculation result.

[0191] Step 405 : When the performance saturation meets the preset saturation condition, the resource allocation strategy is determined as the target allocation strategy.

[0192] Optionally, when the performance saturation is greater than or equal to a preset saturation, the resource allocation strategy is determined as the target allocation strategy. Alternatively, when the performance saturation reaches a target saturation, the resource allocation strategy is determined as the target allocation strategy. The target saturation refers to the maximum performance saturation that can be achieved based on the activity prediction results corresponding to multiple objects under the condition of fixed preset resources.

[0193] The target allocation strategy is used to allocate preset resources among multiple objects.

[0194] In the embodiment of the present application, the target activity can be an activity to recover lost objects in the game.

[0195] Indicative, such as Figure 5 As shown, in the function graph 500 , the horizontal axis is the activity input cost X, and the vertical axis is the benefit after the cost input, that is, the number of objects d that the activity brings back. Figure 5 Taking the estimated number of returning objects that an activity can bring as an example, when the cost investment is positive, that is, costs are invested in users, such as giving users benefits or issuing coupons of a certain amount, the number of returning objects will increase accordingly after the discount.

[0196] In summary, the embodiment of the present application provides a strategy determination method, which obtains the target activity participation probability corresponding to the multiple objects by performing feature analysis on the object data information corresponding to the multiple objects respectively; then, based on the participation probabilities corresponding to the multiple objects respectively, the resource allocation quotas corresponding to the multiple objects respectively are weighted and summed to obtain the efficiency saturation. When the efficiency saturation meets the preset saturation condition, the resource allocation strategy is determined as the target allocation strategy, thereby allocating preset resources among the multiple objects according to the target allocation strategy. In particular, when determining the target allocation strategy, the expected participation probability of the object in the target activity is taken into account, which improves the rationality of the target allocation strategy, avoids the unreasonable phenomenon of allocating too many resources to objects that do not participate in the target activity, and avoids the waste of activity resources.

[0197] In some embodiments, the performance saturation is used to characterize the activity consumption of the target activity when the resource allocation strategy is executed. Figure 2 or Figure 3 The illustrated embodiment may also be implemented as the following steps 601 to 605 .

[0198] Step 601: Obtain object data information corresponding to a plurality of objects.

[0199] The object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period.

[0200] Step 602 : performing feature analysis on the object data information corresponding to the multiple objects to obtain activity consumption probabilities corresponding to the multiple objects.

[0201] The activity consumption probability is used to represent the predicted probability of the subject participating in the target activity and making consumption, and the target activity includes activities to be carried out on the first interactive platform.

[0202] Optionally, feature engineering processing is performed on the object data information corresponding to the multiple objects to obtain object feature representations corresponding to the multiple objects; the object feature representations corresponding to the multiple objects are predicted through a logistic regression model to obtain activity consumption probabilities corresponding to the multiple objects.

[0203] Step 603: Determine the resource allocation quotas corresponding to the multiple objects.

[0204] The resource allocation quota is used to indicate the allocation quota of the preset resources put in by the first interactive platform among the multiple objects. The preset resources include the resource investment quota corresponding to the target activity.

[0205] Step 604 : performing weighted summation on the resource allocation quotas corresponding to the multiple objects based on the activity consumption probabilities corresponding to the multiple objects to obtain performance saturation.

[0206] Among them, efficacy saturation is positively correlated with the number of consumers of the target activity.

[0207] It should be noted that the resource allocation quota can also be implemented as a resource allocation ratio, that is, the resource allocation ratios corresponding to the multiple objects are weighted and summed based on the activity consumption probabilities corresponding to the multiple objects to obtain the performance saturation.

[0208] In some embodiments, exclusive performance saturation calculations may be performed for target-type items in target activities.

[0209] Target items include items provided to the target for consumption during the target activity, such as event-exclusive skins, event treasure chests, and event-exclusive characters.

[0210] Optionally, the performance saturation is determined based on activity consumption probabilities respectively corresponding to the multiple objects, resource allocation strategies, and type weights respectively corresponding to the multiple objects.

[0211] Among them, the type weight is positively correlated with the probability of the subject consuming the target type items, and the efficacy saturation is positively correlated with the predicted consumption quantity corresponding to the target type items in the target activity.

[0212] For example: based on the activity consumption probabilities corresponding to multiple objects and the type weights corresponding to multiple objects, the resource allocation amounts corresponding to multiple objects are weighted and summed to obtain the efficiency saturation; or, based on the activity consumption probabilities corresponding to multiple objects and the type weights corresponding to multiple objects, the resource allocation ratios corresponding to multiple objects are weighted and summed to obtain the efficiency saturation.

[0213] Optionally, the preset resources in the above resource allocation strategy refer to the resource allocation quota corresponding to the target type of items.

[0214] In the above embodiment, when calculating the performance saturation, the probability of the object consuming the target type of items is taken into account. The calculated performance saturation reflects the probability of the object participating in the activity and consuming the target type of items, so that a targeted allocation strategy is formulated for the target type of items. For example: if the allocation amount of object 1 in the final target allocation strategy is 100, then a coupon with an amount of 100 can be given to object 1.

[0215] Step 605 : When the performance saturation meets the preset saturation condition, the resource allocation strategy is determined as the target allocation strategy.

[0216] The target allocation strategy is used to allocate preset resources among multiple objects.

[0217] In summary, the embodiment of the present application provides a strategy determination method, which obtains the target activity consumption probability corresponding to the multiple objects by performing feature analysis on the object data information corresponding to the multiple objects respectively; then, based on the consumption probabilities corresponding to the multiple objects respectively, the resource allocation quotas corresponding to the multiple objects respectively are weighted and summed to obtain the efficiency saturation. When the efficiency saturation meets the preset saturation conditions, the resource allocation strategy is determined as the target allocation strategy, thereby allocating preset resources among the multiple objects according to the target allocation strategy. Among them, when determining the target allocation strategy, the expected consumption probability of the object for the target activity is taken into account, which improves the rationality of the target allocation strategy, avoids the unreasonable phenomenon of allocating too many resources to objects that do not consume, and avoids the waste of activity resources.

[0218] For illustration, please refer to Figure 7 , which shows a structural block diagram of a strategy determination device provided by an exemplary embodiment of the present application, the device includes the following modules:

[0219] An acquisition module 710 is configured to acquire object data information corresponding to a plurality of objects, wherein the object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period;

[0220] An analysis module 720 is configured to perform feature analysis on the object data information corresponding to the plurality of objects to obtain activity prediction results corresponding to the plurality of objects, wherein the activity prediction results are used to represent the predicted participation of the objects in a target activity, wherein the target activity includes an activity to be carried out on the first interactive platform;

[0221] A determination module 730 is configured to determine resource allocation strategies corresponding to the plurality of objects, wherein the resource allocation strategies are configured to indicate allocation of preset resources delivered by the first interactive platform among the plurality of objects;

[0222] The determining module 730 is further configured to determine a performance saturation based on the activity prediction results corresponding to the multiple objects and the resource allocation strategy, wherein the performance saturation is used to represent the predicted benefit corresponding to the target activity when the resource allocation strategy is executed;

[0223] The determining module 730 is configured to determine the resource allocation strategy as a target allocation strategy when the performance saturation satisfies a preset saturation condition, wherein the target allocation strategy is used to allocate the preset resources among the multiple objects.

[0224] Please refer to Figure 8 In some embodiments, the determination module 730 is further used to determine the resource allocation strategy as the target allocation strategy when the performance saturation reaches the target saturation, and the target saturation refers to the maximum performance saturation that can be achieved based on the activity prediction results corresponding to the multiple objects when the preset resources are fixed.

[0225] In some embodiments, the determination module 730 is further configured to determine the resource allocation strategy as the target allocation strategy when the performance saturation is greater than or equal to a preset saturation.

[0226] In some embodiments, the analysis module 720 includes:

[0227] A processing unit 721 is configured to perform feature engineering processing on the object data information corresponding to the plurality of objects to obtain object feature representations corresponding to the plurality of objects; wherein the feature engineering processing includes at least one of missing value repair, discretization processing, and normalization processing;

[0228] The prediction unit 722 is configured to predict the object feature representations corresponding to the plurality of objects respectively through a logistic regression model to obtain activity prediction results corresponding to the plurality of objects respectively.

[0229] In some embodiments, the acquisition module 710 is used to obtain the activity feature representation corresponding to the target activity; the prediction unit 722 is used to predict the object feature representation and the activity feature representation corresponding to the multiple objects respectively through a logistic regression model to obtain the activity prediction results corresponding to the multiple objects respectively.

[0230] In some embodiments, the activity prediction result includes an activity participation probability, which is used to characterize the predicted probability of the object participating in the target activity; the preset resources include the resource investment amount corresponding to the target activity; the determination module 730 is also used to perform weighted summation of the resource allocation amounts corresponding to the multiple objects based on the activity participation probabilities corresponding to the multiple objects, to obtain the performance saturation; wherein, the performance saturation is positively correlated with the number of activity participating objects of the target activity.

[0231] In some embodiments, the activity prediction result further includes a predicted activity level, where the activity level is used to represent the predicted activity level of the subject in the target activity. The determination module 730 includes:

[0232] a grouping unit 731 configured to group the plurality of objects according to m activity intervals to obtain m groups of objects, wherein the predicted activity of the i-th group of objects is within the i-th activity interval, where m is an integer greater than 1, i≤m, and i is a positive integer;

[0233] A determining unit 732 is configured to determine activity weights corresponding to the m groups of objects, wherein the activity weights are positively correlated with activity levels corresponding to the activity intervals;

[0234] The determining unit 732 is further configured to perform weighted summation of the resource allocation quotas corresponding to the multiple objects based on the activity participation probabilities corresponding to the multiple objects and the activity weights corresponding to the m groups of objects to obtain the performance saturation.

[0235] In some embodiments, the activity prediction result includes an activity consumption probability, which is used to characterize the predicted probability of the object participating in the target activity and making consumption; the determination module 730 is used to determine the performance saturation based on the activity consumption probabilities corresponding to the multiple objects, the resource allocation strategy, and the type weights corresponding to the multiple objects; wherein the type weight is positively correlated with the probability of the object consuming target type items, and the target type items include items provided to the object for consumption in the target activity; the performance saturation is positively correlated with the predicted consumption quantity corresponding to the target type items in the target activity.

[0236] In some embodiments, the apparatus comprises:

[0237] An allocation module 740 is configured to allocate target resources to a target object according to a resource allocation result of the target object in the target allocation strategy when the target activity is started and the target object among the multiple objects logs into the first interactive platform;

[0238] The target resources include at least one of activity resources, physical resources, and promotional resources. The activity resources include resources provided to the object for use in the target activity, and the promotional resources include resources provided to the object for use in other interactive platforms.

[0239] In some embodiments, the allocation module 740 is also used to deliver the target resources to the designated object according to the resource allocation result of the target object in the target allocation strategy when the target activity is started and the target object has not logged into the first interactive platform within a preset time period; wherein the preset time period is a time period starting at the start time of the target activity, and the duration corresponding to the preset time period is less than or equal to the activity duration of the target activity; the designated object refers to an object whose activity in the target activity is greater than or equal to the preset activity.

[0240] In summary, the strategy determination device provided in the embodiment of the present application performs feature analysis on the object data information corresponding to the multiple objects respectively, and obtains the target activity participation status corresponding to the multiple objects respectively; then, based on the participation status corresponding to the multiple objects and the determined resource allocation strategy, the performance saturation is determined, and when the performance saturation meets the preset saturation condition, the resource allocation strategy is determined as the target allocation strategy, so that the preset resources are allocated among the multiple objects according to the target allocation strategy. Based on the activity prediction results and the resource allocation strategy, the performance saturation used to measure the predicted benefits is determined, and the target allocation strategy is determined under the constraints of the preset saturation condition, so that the activity resources can be concentrated on the objects with higher expected benefits, ensuring that the activity resources are maximized and avoiding the waste of activity resources.

[0241] It should be noted that the policy determination device provided in the above embodiment is merely an example of the division of the aforementioned functional modules. In actual applications, the aforementioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the policy determination device provided in the above embodiment and the policy determination method embodiment are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0242] Figure 9The following is a block diagram of a computer device 900 according to an exemplary embodiment of the present application. The computer device 900 may be a smartphone, a tablet computer, a Moving Picture Experts Group Audio Layer III (MP3) player, a Moving Picture Experts Group Audio Layer IV (MP4) player, a laptop computer, or a desktop computer. The computer device 900 may also be referred to as a user device, a portable computer device, a laptop computer device, a desktop computer device, or other similar names.

[0243] Typically, the computer device 900 includes a processor 901 and a memory 902 .

[0244] The processor 901 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 901 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 901 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a central processing unit (CPU); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 901 may be integrated with a graphics processing unit (GPU), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 901 may also include an artificial intelligence (AI) processor, which is used to process computing operations related to machine learning.

[0245] Memory 902 may include one or more computer-readable storage media, which may be non-transitory. Memory 902 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in memory 902 is used to store at least one instruction, which is executed by processor 901 to implement the virtual game display method provided in the method embodiment of the present application.

[0246] Schematically, the computer device 900 also includes other components, which can be understood by those skilled in the art. Figure 9 The structure shown in the figure does not constitute a limitation on the computer device 900, and the computer device 900 may include more or fewer components than shown in the figure, or combine some components, or adopt a different component arrangement.

[0247] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments can be performed by instructing the relevant hardware through a program. The program can be stored in a computer-readable storage medium, which can be the computer-readable storage medium included in the memory of the above embodiments, or a separate computer-readable storage medium not incorporated into the computer device. The computer-readable storage medium stores at least one instruction, at least one program, code set, or instruction set. The at least one instruction, at least one program, code set, or instruction set is loaded and executed by the processor to implement the virtual game display method described in any of the above embodiments.

[0248] Optionally, the computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a solid-state drive (SSD), or an optical disk. Among them, the random access memory may include a resistance random access memory (ReRAM) and a dynamic random access memory (DRAM). The serial numbers of the above embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0249] Those skilled in the art will appreciate that all or part of the steps in the above embodiments may be implemented by hardware or by programs instructing the relevant hardware to perform the steps. The programs may be stored in a computer-readable storage medium, such as a read-only memory, a magnetic disk, or an optical disk. The above are merely optional embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A strategy determination method, characterized in that: The method comprises: Obtaining object data information corresponding to a plurality of objects, wherein the object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period; Performing feature analysis on object data information corresponding to each of the plurality of objects to obtain activity prediction results corresponding to each of the plurality of objects, the activity prediction results being used to characterize the predicted participation of the objects in a target activity, the target activity comprising an activity to be carried out on the first interactive platform; Determining resource allocation strategies corresponding to the multiple objects, the resource allocation strategies being used to indicate allocation of preset resources delivered by the first interactive platform among the multiple objects; Determining, based on the activity prediction results corresponding to the plurality of objects and the resource allocation strategy, a performance saturation, wherein the performance saturation is used to represent the predicted benefit corresponding to the target activity when the resource allocation strategy is executed; In a case where the performance saturation meets a preset saturation condition, the resource allocation strategy is determined as a target allocation strategy, and the target allocation strategy is used to allocate the preset resources among the multiple objects.

2. The method according to claim 1, characterized in that When the performance saturation meets the preset saturation condition, determining the resource allocation strategy as the target allocation strategy includes: When the performance saturation reaches the target saturation, the resource allocation strategy is determined as the target allocation strategy. The target saturation refers to the maximum performance saturation that can be achieved based on the activity prediction results corresponding to the multiple objects when the preset resources are fixed.

3. The method according to claim 1, characterized in that When the performance saturation meets the preset saturation condition, determining the resource allocation strategy as the target allocation strategy includes: When the performance saturation is greater than or equal to a preset saturation, the resource allocation strategy is determined as the target allocation strategy.

4. The method according to any one of claims 1 to 3, characterized in that: The performing feature analysis on the object data information corresponding to the plurality of objects to obtain activity prediction results corresponding to the plurality of objects includes: Performing feature engineering processing on the object data information corresponding to the plurality of objects to obtain object feature representations corresponding to the plurality of objects; wherein the feature engineering processing includes at least one of missing value repair, discretization processing, and normalization processing; The object feature representations respectively corresponding to the multiple objects are predicted using a logistic regression model to obtain activity prediction results respectively corresponding to the multiple objects.

5. The method according to claim 4, characterized in that The method further comprises: Obtaining an activity feature representation corresponding to the target activity; The predicting the object feature representations corresponding to the plurality of objects respectively by using a logistic regression model to obtain activity prediction results corresponding to the plurality of objects respectively includes: The object feature representations and the activity feature representations respectively corresponding to the multiple objects are predicted using a logistic regression model to obtain activity prediction results respectively corresponding to the multiple objects.

6. The method according to any one of claims 1 to 3, characterized in that: The activity prediction result includes an activity participation probability, which is used to represent the predicted probability of the subject participating in the target activity; the preset resources include the resource investment amount corresponding to the target activity; The determining of the performance saturation based on the activity prediction results corresponding to the plurality of objects and the resource allocation strategy includes: Performing weighted summation of resource allocation quotas corresponding to the multiple objects based on activity participation probabilities corresponding to the multiple objects, to obtain the performance saturation; The effectiveness saturation is positively correlated with the number of participants in the target activity.

7. The method according to claim 6, characterized in that The activity prediction result also includes a predicted activity level, where the activity level is used to represent the predicted activity level of the subject in the target activity; The step of performing weighted summation of the resource allocation quotas corresponding to the multiple objects based on the activity participation probabilities corresponding to the multiple objects to obtain the performance saturation includes: Grouping the plurality of objects according to m activity intervals to obtain m groups of objects, wherein the predicted activity of the i-th group of objects is in the i-th activity interval, where m is an integer greater than 1, i≤m, and i is a positive integer; Determining activity weights corresponding to the m groups of objects, respectively, where the activity weights are positively correlated with the activity levels corresponding to the activity intervals; The performance saturation is obtained by performing a weighted summation on the resource allocation quotas corresponding to the multiple objects based on the activity participation probabilities corresponding to the multiple objects and the activity weights corresponding to the m groups of objects.

8. The method according to any one of claims 1 to 3, characterized in that: The activity prediction result includes an activity consumption probability, which is used to represent the predicted probability of the subject participating in the target activity and consuming; The determining of the performance saturation based on the activity prediction results corresponding to the plurality of objects and the resource allocation strategy includes: determining the performance saturation based on activity consumption probabilities respectively corresponding to the plurality of objects, the resource allocation strategy, and type weights respectively corresponding to the plurality of objects; Among them, the type weight is positively correlated with the probability of the object consuming the target type items, and the target type items include items provided to the object for consumption in the target activity; the performance saturation is positively correlated with the predicted consumption quantity corresponding to the target type items in the target activity.

9. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: When the target activity is started and a target object among the multiple objects logs in to the first interactive platform, delivering target resources to the target object according to the resource allocation result of the target object in the target allocation strategy; The target resources include at least one of activity resources, physical resources, and promotional resources. The activity resources include resources provided to the object for use in the target activity, and the promotional resources include resources provided to the object for use in other interactive platforms.

10. The method according to claim 9, characterized in that The method further comprises: When the target activity is started and the target object has not logged into the first interactive platform within a preset time period, delivering the target resource to the designated object according to the resource allocation result of the target object in the target allocation strategy; Among them, the preset time period is a time period starting at the start time of the target activity, and the duration corresponding to the preset time period is less than or equal to the activity duration of the target activity; the designated object refers to an object whose activity in the target activity is greater than or equal to the preset activity.

11. A strategy determination device, characterized in that: The device comprises: An acquisition module, configured to acquire object data information corresponding to a plurality of objects, wherein the object data information includes historical activity data, and the historical activity data is used to indicate the object's participation in activities on the first interactive platform within a historical time period; an analysis module, configured to perform feature analysis on the object data information corresponding to the plurality of objects, to obtain activity prediction results corresponding to the plurality of objects, the activity prediction results being used to characterize the predicted participation of the objects in a target activity, the target activity comprising an activity to be carried out on the first interactive platform; a determination module, configured to determine resource allocation strategies corresponding to the plurality of objects, wherein the resource allocation strategies are configured to indicate allocation of preset resources delivered by the first interactive platform among the plurality of objects; The determining module is further configured to determine a performance saturation based on the activity prediction results corresponding to the plurality of objects and the resource allocation strategy, wherein the performance saturation is used to represent the predicted benefit corresponding to the target activity when the resource allocation strategy is executed; The determining module is configured to determine the resource allocation strategy as a target allocation strategy when the performance saturation satisfies a preset saturation condition, wherein the target allocation strategy is used to allocate the preset resources among the multiple objects.

12. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one program, and the at least one program is loaded and executed by the processor to implement the policy determination method according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that The storage medium stores at least one program, and the at least one program is loaded and executed by the processor to implement the policy determination method according to any one of claims 1 to 10.

14. A computer program product, characterized in that The method comprises a computer program, which implements the policy determination method according to any one of claims 1 to 10 when the computer program is executed by a processor.

Images