Methods for determining excitation duration and related equipment

Abstract

This application discloses a method and related equipment for determining the incentive duration. The method includes: acquiring historical data of a target object in response to a campaign message during a historical period, the historical incentive duration corresponding to the target object in the historical period, and the characteristic information of the target object; predicting the estimated information exposure parameters and estimated information click parameters of the target object in a future target period under each candidate incentive duration based on the historical data, historical incentive duration, and characteristic information of the target object; determining the target incentive duration from at least two candidate incentive durations based on the estimated information exposure parameters and information click parameters of the target object in the future target period under each candidate incentive duration, so that if the target object's interaction time with the campaign message reaches the target incentive duration in the target period, virtual resources are distributed to the target object. This solution can improve the reach rate of campaign messages.

Description

Technical Field

[0001] This application relates to the field of computer technology, and more specifically, to a method and related equipment for determining the duration of an excitation. Background Technology

[0002] To improve the efficiency of information acquisition for users, information is delivered to users with high relevance to their needs, thereby reducing the time users spend filtering and selecting information. However, related technologies suffer from low reach rates during the delivery process, leading to significant network resource consumption and low network resource utilization efficiency. Summary of the Invention

[0003] In view of the above problems, embodiments of this application propose a method and related equipment for determining the excitation duration to improve the above problems.

[0004] In a first aspect, this application provides a method for determining the incentive duration, comprising: acquiring target historical data of a target object in response to information delivery in a historical period, the historical incentive duration corresponding to the target object in the historical period, and the characteristic information of the target object; predicting, based on the target historical data, the historical incentive duration, and the characteristic information of the target object, the estimated information exposure parameters and estimated information click parameters corresponding to the target object in a future target period under each candidate incentive duration; determining a target incentive duration from at least two candidate incentive durations based on the estimated information exposure parameters and information click parameters corresponding to the target object in a future target period under each candidate incentive duration, wherein, if the interaction duration of the target object in response to the information delivery reaches the target incentive duration in the target period, virtual resources are distributed to the target object.

[0005] Secondly, this application provides a method for distributing virtual resources, the method comprising: displaying distribution information; responding to a click operation triggered by the distribution information, calculating the interaction duration of an object with respect to the distribution information; if the interaction duration reaches the incentive duration corresponding to the object in the current period, then distributing virtual resources to the object, wherein the incentive duration corresponding to the object in the current period is determined according to the above-described method for determining incentive duration.

[0006] Thirdly, this application provides an apparatus for determining the duration of an incentive, comprising: an acquisition module, configured to acquire target historical data of a target object in response to information delivery in a historical period, the historical incentive duration of the target object in the historical period, and the characteristic information of the target object; a prediction module, configured to predict, based on the target historical data, the historical incentive duration, and the characteristic information of the target object, the estimated information exposure parameters and estimated information click parameters of the target object in a future target period under each candidate incentive duration; and a target incentive duration determination module, configured to determine a target incentive duration from at least two candidate incentive durations based on the estimated information exposure parameters and information click parameters of the target object in a future target period under each candidate incentive duration, so that if the interaction duration of the target object in response to the information delivery reaches the target incentive duration in the target period, virtual resources are distributed to the target object.

[0007] In some embodiments, the target incentive duration determination module includes: a first weighting coefficient acquisition unit, configured to acquire a first weighting coefficient corresponding to the estimated information exposure parameter; a second weighting coefficient acquisition unit, configured to acquire a second weighting coefficient corresponding to the estimated information click parameter; a first weighting processing unit, configured to perform weighting processing on the estimated information exposure parameter and the estimated information click parameter corresponding to the target object in the future target period under each candidate incentive duration, based on the first weighting coefficient corresponding to the estimated information exposure parameter and the second weighting coefficient corresponding to the estimated information click parameter, to obtain a weighted result corresponding to each candidate incentive duration; and a target incentive duration determination unit, configured to determine the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

[0008] In other embodiments, the target excitation duration determination module includes: an exposure change parameter calculation unit, configured to calculate, based on the estimated information exposure parameters corresponding to the target object under each candidate excitation duration, an exposure change parameter of the target object relative to a reference excitation duration, wherein the reference excitation duration is one of at least two candidate excitation durations; a click change parameter calculation unit, configured to calculate, based on the estimated information click parameters corresponding to the target object under each candidate excitation duration, a click change parameter of the target object relative to a reference excitation duration; and a first weighting coefficient acquisition unit. The system comprises: a first weighting coefficient acquisition unit for obtaining a first weighting coefficient corresponding to the estimated information exposure parameter; a second weighting coefficient acquisition unit for obtaining a second weighting coefficient corresponding to the estimated information click parameter; a second weighting processing unit for weighting the exposure change parameter and click change parameter corresponding to each candidate incentive duration based on the first weighting coefficient and the second weighting coefficient corresponding to the estimated information click parameter, to obtain a weighted result corresponding to each candidate incentive duration; and a target incentive duration determination unit for determining the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

[0009] In some embodiments, the second weighting coefficient acquisition unit includes: an object layer determination unit, configured to determine the target object layer to which the target object belongs; and a coefficient determination unit, configured to determine the weighting coefficient corresponding to the target object layer based on the mapping relationship between the object layer and the weighting coefficient, and use the weighting coefficient corresponding to the target object layer as the second weighting coefficient.

[0010] In some embodiments, the target historical data includes information identifiers of each delivery message exposed to the target object during the historical period; in this embodiment, the object stratification determination unit includes: an exposure count statistics unit, used to count the exposure counts corresponding to each delivery message to the target object during the historical period based on the target historical data; a revenue unit price parameter acquisition unit, used to acquire the revenue parameters corresponding to each delivery message exposed to the target object during the historical period based on the information identifiers included in the target historical data; a target total revenue parameter calculation unit, used to calculate the target revenue parameter corresponding to the target object during the historical period based on the exposure counts corresponding to each delivery message to the target object during the historical period and the revenue parameters corresponding to each delivery message exposed to the target object during the historical period; and a determination unit, used to determine the object stratification corresponding to the target total revenue parameter as the target object stratification to which the target object belongs.

[0011] In some embodiments, the prediction module includes: a combination unit, configured to combine each candidate excitation duration with the target historical data, the historical excitation duration, and the feature information of the target object to obtain at least two sets of input information; a first prediction unit, configured to use an exposure prediction model to predict exposure parameters based on each set of input information, and output the predicted exposure parameters of the target object in the future target period under each candidate excitation duration; and a second prediction unit, configured to use a click prediction model to predict click parameters based on each set of input information, and output the predicted click parameters of the target object in the future target period under each candidate excitation duration.

[0012] In some embodiments, the device for determining the incentive duration further includes: a training data acquisition module, configured to acquire training data, the training data including multiple training samples, the training samples including at least two candidate incentive durations corresponding to the sample user in a first historical period, first target historical data of the sample user for the delivery information under each candidate incentive duration in the first historical period, feature information of the sample user, and sample incentive duration corresponding to the sample user in a second historical period; wherein, in terms of time sequence, the second historical period is later than the first historical period; the training module is configured to back-train the exposure prediction model and the click prediction model based on the at least two candidate incentive durations corresponding to the sample user in the first historical period, the first target historical data of the sample user for the delivery information under each candidate incentive duration in the first historical period, the feature information of the sample user, and the sample incentive duration corresponding to the sample user in the second historical period, until the training termination condition is reached.

[0013] Fourthly, this application provides a virtual resource distribution device, comprising: a display module for displaying distribution information; a statistics module for counting the interaction duration of an object in response to a click operation triggered by the distribution information; and a distribution module for distributing virtual resources to the object if the interaction duration reaches the incentive duration corresponding to the object in the current period, wherein the incentive duration corresponding to the object in the current period is determined according to the above-described method for determining incentive duration.

[0014] In some embodiments, the virtual resource distribution device further includes: a reporting module, configured to upload the exposure record of the distribution information on the client where the object is located, the click record of the object on the distribution information, and the total interaction duration of the object on the distribution information to the server, so that the server can predict the incentive duration of the object in the next cycle.

[0015] Fifthly, an electronic device is provided, comprising: a processor; and a memory storing computer-readable instructions, wherein when the computer-readable instructions are executed by the processor, the method for determining the incentive duration or the method for distributing virtual resources as described above is implemented.

[0016] Sixthly, according to one aspect of the embodiments of this application, a computer-readable storage medium is provided, on which computer-readable instructions are stored, which, when executed by a processor, implement the method for determining the incentive duration or the method for distributing virtual resources as described above.

[0017] Seventhly, according to one aspect of the embodiments of this application, a computer program product is provided, including computer instructions, which, when executed by a processor, implement the above-described method for determining the duration of the incentive or the method for distributing virtual resources.

[0018] In this application, historical data on the target audience's response to the delivered information, along with the target audience's historical incentive duration and characteristic information, are used to predict the estimated information exposure parameters and estimated information click parameters for the target audience under each candidate incentive duration. Since these parameters reflect the target audience's level of attention and interest in the delivered information at each candidate incentive duration, the inventors recognized that setting the incentive duration too long could negatively impact user experience and reduce the reach of the delivered information. Furthermore, different audiences exhibit varying levels of attention and interest in the delivered information even under the same candidate incentive duration. Therefore, this application utilizes these two parameters—the estimated information exposure parameters and the estimated information click parameters—to specifically determine the target incentive duration for the target audience in future target periods. This approach improves the reach of the delivered information and, by specifically determining an appropriate target incentive duration for the target audience, ensures a better overall user experience. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0020] Figure 1 This is a schematic diagram illustrating an application scenario of the present solution according to an embodiment of this application.

[0021] Figure 2 This is a flowchart illustrating a method for determining the excitation duration according to an embodiment of this application.

[0022] Figure 3 This is a flowchart illustrating the process of determining the target object hierarchy to which a target object belongs, according to an embodiment of this application.

[0023] Figure 4 This is a flowchart illustrating step 220 according to an embodiment of this application.

[0024] Figure 5 This is a flowchart illustrating a method for distributing virtual resources according to an embodiment of this application.

[0025] Figure 6 The diagram illustrates the changes in ad exposure decrease and revenue increase as a function of incentive duration.

[0026] Figure 7A This is a schematic diagram illustrating the training of an exposure prediction model and a click prediction model based on a specific implementation.

[0027] Figure 7B This is a flowchart illustrating the periodic determination of excitation duration according to a specific embodiment of this application.

[0028] Figure 8 This is a block diagram of an excitation duration determination device according to an embodiment of this application.

[0029] Figure 9 This is a block diagram of a virtual resource distribution device according to an embodiment of this application.

[0030] Figure 10 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0031] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art.

[0032] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0033] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0034] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0035] It should be noted that "multiple" in this article refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0036] Figure 1 This is a schematic diagram illustrating an application scenario of the solution according to an embodiment of this application. For example... Figure 1 As shown, this application scenario can include terminal 110 and server 120. Terminal 110 and server 120 can establish direct or indirect communication connections through wired or wireless networks. Server 120 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0037] Terminal 110 can be a smartphone, tablet, laptop, desktop computer, in-vehicle terminal, smart TV, or other electronic device that can interact with the user. Terminal 110 can run applications, which can be news client programs, interactive aggregation platform client programs, game client programs, video client programs, music client programs, reading application client programs, browser client programs, or other client programs that can display and deliver information.

[0038] The server 120 can be used to execute the method of this application, determine the target incentive duration corresponding to each target user in the future target period according to the method of this application, and send the determined target incentive duration to the terminal 110.

[0039] The server 120 can also be used to send delivery information to the terminal 110. After receiving and displaying the delivery information from the server 120, the application running on the terminal 110 can detect the user's interaction duration with the delivery information. Once the interaction duration reaches the user's target incentive duration for the current period, virtual resources are distributed to the user. It can be understood that the virtual resources distributed to the user are essentially distributed to the user's account on the application.

[0040] In some embodiments, after the application on terminal 110 detects that the user's interaction time with the delivered information has reached the target incentive time for the user in the current period, it can send a notification message to server 120. Upon receiving the notification message, server 120 will issue virtual resources to the user. The virtual resources may be one or more of the following: lottery opportunities, coupons, feature unlocks, account upgrades, etc., without specific limitations.

[0041] Furthermore, after detecting interactive behavior in response to the delivery information, the application in terminal 110 generates a corresponding interactive behavior record and sends the interactive behavior record to server 120 so that server 120 can use the interactive behavior record to determine the target incentive duration for the user in the next cycle.

[0042] In some embodiments, the server 120 may also deploy a database to store delivery information. After retrieving delivery information from the database, the server sends the delivery information to the user's terminal 110. In some embodiments, the server 120 may also classify the delivery information to be delivered by user based on the delivery information to be delivered in the database, thereby obtaining a delivery information set corresponding to each user. When it is necessary to send delivery information to a user's terminal, the server retrieves the delivery information from the delivery information set corresponding to that user.

[0043] In some embodiments, the delivery information in this application can be embedded in other main content, so that the delivery information is displayed correspondingly during the display of the main content. In this case, the application in the terminal can also determine the content category to which the main content displayed on the terminal's display interface belongs; then send the content category to which the main content belongs to the server 120, and the server 120 selects delivery information that matches the content category from the database according to the content category, and delivers the selected delivery information to the terminal 110, thereby achieving matching between the category of the main content displayed on the terminal 110 and the category of the delivery information.

[0044] Taking advertising as an example, if the main content displayed on terminal 110 is a game video, its corresponding content category could be the game category, and the selected advertising information could be a game advertisement. If the main content displayed on terminal 110 is a travelogue article, its corresponding content category could be the travel category, and the selected advertising information could be a travel package advertisement. Because the advertising information to be delivered is selected based on the content category of the main content displayed on the terminal, the delivery and exposure of the advertising information can be more targeted.

[0045] The implementation details of the technical solutions in the embodiments of this application are described in detail below:

[0046] Figure 2 This is a flowchart illustrating a method for determining the excitation duration according to an embodiment of this application. This method can be executed by an electronic device with processing capabilities, such as… Figure 1 The server and terminal components mentioned are not specifically defined here. (See reference...) Figure 2 As shown, the method includes at least steps 210 to 230, which are described in detail below:

[0047] Step 210: Obtain the target historical data of the target object in the historical period for the information delivery, the historical incentive duration of the target object in the historical period, and the characteristic information of the target object.

[0048] The information to be distributed can be information that needs to be promoted / distributed, such as advertisements, announcements / notifications, promotional information for public welfare activities, news, discussion topics, blogs, articles on public accounts, live videos, recorded videos, product links, etc. The form of the information to be distributed can be video, text, audio and video, audio, images, etc., without specific limitations.

[0049] The target object generally refers to the object whose incentive duration needs to be determined. In this application, the object can refer to a user registered in the application. Specifically, the object can be distinguished by user identifiers in the application (such as user account, user ID, etc.).

[0050] The historical period refers to the period during which information dissemination and promotion have been carried out. The historical period can be one period or multiple periods. In a specific embodiment, to ensure the timeliness of the target historical data relative to the future target period, the historical period can be the N periods preceding the target period, where N is a positive integer. When N is 1, the historical period is the period preceding the target period. It is understood that in this application, the target period is later than the historical period in terms of time sequence. In a specific embodiment, the duration of a period can be set according to actual needs, such as a week, a day, a month, etc., without specific limitations here.

[0051] Target historical data is determined based on the interactive behaviors of the target object in response to previously delivered ad messages within a historical period. Target historical data can include records of the target object's interactive behaviors in response to several previously delivered ad messages during the historical period. During the historical period, ad messages have already been sent to the client where the target object resides. Subsequently, corresponding interactive behavior records are generated based on the interactive behaviors triggered by the target object in response to the ad messages. These interactive behavior records can be determined based on logs generated by the client where the target object resides. For example, if the client where the target object resides triggers a click action, the client generates a corresponding click log, which can be used as a click action record.

[0052] Interactive behaviors triggered by the target audience in response to the information can include exposure of the information, click behavior, video playback time reaching a first specified duration, dwell time on the page of the information (which can be understood as browsing time) reaching a second specified duration, conversion behavior (such as purchase behavior triggered by product links, purchase behavior triggered by product links in advertisements, etc.), sharing behavior (or forwarding behavior), liking behavior, posting comments, etc., without specific limitations.

[0053] In some embodiments, since interaction behavior records for certain interactive behaviors may not be very useful in determining the incentive duration, in order to balance data processing efficiency, the interaction behavior records corresponding to specific interactive behaviors can be selected as the data basis for determining the incentive duration. For example, if the specified interactive behaviors include exposure behaviors and click behaviors, then the target historical data includes exposure behavior records and click behavior records. For example, if the specified interactive behaviors include exposure behaviors, click behaviors, and sharing behaviors, then the target historical data includes exposure behavior records, click behavior records, and sharing behavior records.

[0054] Interaction behavior records can indicate triggered interactions, such as exposure or clicks. It's understood that triggering an exposure action means entering the page containing the ad content.

[0055] In some embodiments, interaction behavior records can also indicate the exposure location of the delivery information on the client of the target audience. For example, if the delivery information is an advertisement, the exposure location of the advertisement can be the application's cover page, the start of a video, a certain playback position during video playback, the end playback position of the video, the top position of the page, the middle of the page, the bottom of the page, etc. Specifically, the location that can be used to display the delivery information can be pre-specified, so that the delivery information is displayed in the specified location when it is obtained.

[0056] In some embodiments, since the same object may share an object account in multiple applications, such as an object sharing an account in three applications: instant messaging application I, instant messaging application II, and game application A, or an object authorizing to log in in instant messaging application II and game application A using the account in instant messaging application I, in this case, the interaction behavior record may also include the application identifier corresponding to the application that displays the delivery information.

[0057] Historical incentive duration refers to the incentive duration corresponding to the target object within a historical period. In the scheme of this application, the incentive duration is determined for each object on a periodic basis, and the incentive duration is updated periodically. It can be understood that if there are multiple historical periods, then the corresponding historical incentive duration includes the incentive duration corresponding to each period of the target object within the historical period.

[0058] The characteristic information can be one or more of the following: gender, age, location, marital status, interest / preference tags, account level, membership status (whether it is a member, the membership level, etc.), without specific limitations.

[0059] It is understood that in the specific implementation of this application, data related to user information (such as gender, age, location, marital status, interest tags / preference tags, etc. mentioned above) are involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0060] Step 220: Based on the target's historical data, historical incentive duration, and the target object's feature information, predict the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration.

[0061] Estimated information exposure parameters refer to the predicted information exposure parameters, which are parameters used to reflect the exposure status of the delivered information. In specific embodiments, information exposure parameters can be information exposure volume or information exposure rate. Correspondingly, estimated information exposure parameters can be at least one of estimated information exposure volume and estimated information exposure rate. Furthermore, information exposure volume can be the average information exposure volume over a period of time or the total information exposure volume over a period of time.

[0062] Estimated click parameters refer to the predicted click parameters for the delivered information. Click parameters are parameters used to reflect the click situation of the delivered information. In specific embodiments, click parameters can be the number of clicks or the click-through rate (CTR); correspondingly, estimated click parameters can be at least one of estimated click volume and estimated CTR. Furthermore, the number of clicks can be the average number of clicks over a period or the total number of clicks over a period.

[0063] The candidate incentive duration is preset, and in this application, there are multiple candidate incentive durations. These multiple candidate incentive durations, set for different objects, can be the same or different.

[0064] In some embodiments, objects can be categorized, thereby allowing for the targeted setting of candidate incentive durations for each object category.

[0065] In some embodiments, since objects have different levels of acceptance of interaction duration with different forms of information delivery, they are classified according to the preferred information presentation format. For example, objects whose preferred information presentation format is video belong to the first object category, objects whose preferred information presentation format is text belong to the second object category, and objects whose preferred information presentation format is image belong to the third object category. Multiple candidate incentive durations are set for the first object category, the second object category, and the third object category respectively.

[0066] In some embodiments, objects can be categorized based on their interactive behaviors triggered by previously delivered information during a historical period. Specifically, at least one of the following can be calculated based on the target historical data of the object in response to the delivered information during a historical period: exposure, clicks, and interaction duration (wherein, for video-type delivered information, interaction duration can be video playback duration, and for text or image-type delivered information, interaction duration can be viewing duration or dwell time on the page). At least one of the following ranges can be set for each object category: exposure range, click range, and interaction duration range, and multiple candidate incentive durations can be set for each object category.

[0067] For example, in a specific embodiment, the exposure range and click range corresponding to each object category, as well as the multiple candidate incentive durations corresponding to each object category, can be as shown in Table 1 below:

[0068] Table 1

[0069]

[0070] In some embodiments, an exposure prediction model for predicting information exposure parameters and a click prediction model for predicting information click parameters can be constructed respectively. Historical target data, historical excitation duration, and feature information of the target object, along with the duration of each candidate excitation, are input into the exposure prediction model, which outputs the predicted information exposure parameters of the target object in the future target period. Similarly, historical target data, historical excitation duration, and feature information of the target object, along with the duration of each candidate excitation, are input into the click prediction model, which outputs the predicted information click parameters of the target object in the target period.

[0071] In some embodiments, a predictive model can also be constructed to estimate information exposure parameters and information click parameters. The predictive model outputs the estimated information exposure parameters and estimated information click parameters of the target object in the target period based on the target's historical data, historical excitation duration, and the characteristic information of the target object and the duration of each candidate excitation.

[0072] Understandably, in order to ensure the accuracy of the exposure prediction model, click prediction model, and prediction model, it is necessary to train these models in advance so that they can be applied online.

[0073] The aforementioned exposure prediction model, click prediction model, and prediction model can be models built based on classifiers (such as GBDT model (Gradient Boosting Decision Tree), XGBoost model, etc.), or deep learning models built from one or more of recurrent neural networks, convolutional neural networks, long short-term memory neural networks, fully connected neural networks, feedforward neural networks, etc.

[0074] It is understandable that different audiences have different levels of attention and interest in the information they receive. Therefore, the estimated information exposure parameters and estimated information click parameters may differ for different audiences.

[0075] Furthermore, different incentive durations may affect the recipient's attention to subsequent ad placements. For example, if the incentive duration for a particular ad placement is set to be long, the recipient may not pay attention to or pay less attention to several subsequent ad placements.

[0076] Historical data reflects the target audience's interaction with the campaign information over a historical period. This interaction reflects, to some extent, the audience's level of attention to the campaign information, and the change in audience attention to the campaign information is relatively stable over a period of time. Therefore, in this application's solution, considering the impact of the audience itself, the incentive duration, and the audience's corresponding historical data on the audience's level of attention to the campaign information during the target period, the estimated information exposure parameters and estimated information click parameters of the target audience under each candidate incentive duration are predicted by combining the target audience's feature information, at least two candidate incentive durations, and the target audience's historical data on the campaign information over historical periods.

[0077] Step 230: Based on the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration, determine the target incentive duration from at least two candidate incentive durations. If the target object's interaction time with the delivered information reaches the target incentive duration in the target period, virtual resources are distributed to the target object.

[0078] The estimated information exposure parameters and estimated information click parameters for the target audience under each candidate incentive duration reflect the exposure of the delivered information to the target audience and the target audience's click behavior on the delivered information under that candidate incentive duration. Therefore, the estimated information exposure parameters and information click parameters can comprehensively reflect the target audience's attention and interest in the delivered information under the corresponding incentive duration. Thus, by combining the estimated information exposure parameters and information click parameters for the target audience under each candidate incentive duration, the corresponding target incentive duration for the target audience can be comprehensively determined.

[0079] For the target audience, setting an excessively long incentive duration for a previously displayed message may reduce their attention to subsequent messages, thus impacting the overall reach of the message throughout the campaign and potentially degrading the user experience. Therefore, this application's solution comprehensively determines the target incentive duration by considering the target audience's attention to the message across various candidate incentive durations, based on the estimated message exposure and click parameters. This strikes a balance between improving message reach and ensuring a positive user experience.

[0080] In some embodiments, step 230 includes: obtaining a first weighting coefficient corresponding to the estimated information exposure parameter; and obtaining a second weighting coefficient corresponding to the estimated information click parameter; based on the first weighting coefficient corresponding to the estimated information exposure parameter and the second weighting coefficient corresponding to the estimated information click parameter, weighting the estimated information exposure parameter and the estimated information click parameter corresponding to the target object in the future target period under each candidate incentive duration to obtain a weighted result corresponding to each candidate incentive duration; and determining the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

[0081] The weighted processing can be described by the following formula 1:

[0082] ;(Formula 1)

[0083] in, This refers to the weighted result corresponding to the i-th candidate incentive duration; The first weighting coefficient; The estimated information exposure parameters for the target object at the i-th candidate excitation duration; This is the second weighting coefficient; Click the parameter to obtain the estimated information corresponding to the i-th candidate stimulus duration for the target object.

[0084] In some embodiments, step 230 includes: calculating the exposure change parameters of the target object at each candidate incentive duration relative to a reference incentive duration, based on the estimated information exposure parameters corresponding to the target object at each candidate incentive duration, wherein the reference incentive duration is one of at least two candidate incentive durations; calculating the click change parameters of the target object at each candidate incentive duration relative to the reference incentive duration, based on the estimated information click parameters corresponding to the target object at each candidate incentive duration; obtaining a first weighting coefficient corresponding to the estimated information exposure parameters and a second weighting coefficient corresponding to the estimated information click parameters; weighting the exposure change parameters and click change parameters corresponding to each candidate incentive duration based on the first weighting coefficient and the second weighting coefficient corresponding to the estimated information click parameters to obtain a weighted result corresponding to each candidate incentive duration; and determining the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

[0085] Typically, the exposure and click rates of targeted advertising messages differ significantly. For example, exposure usually exceeds clicks. Therefore, to avoid large discrepancies in the weighted results due to significant differences in the estimated exposure and click parameters, the estimated exposure and click parameters can be transformed before weighting to bring them within a similar range of values.

[0086] The exposure variation parameter reflects the difference between the estimated information exposure parameter at the candidate excitation duration and the estimated information exposure parameter at the reference excitation duration. This exposure variation parameter can be the absolute difference between the estimated information exposure parameter at the candidate excitation duration and the estimated information exposure parameter at the reference excitation duration, or it can be a relative difference.

[0087] That is, exposure variation parameters It can be calculated using the following formula 2:

[0088] ;(Formula 2)

[0089] Alternatively, it can be calculated using the following formula 3:

[0090] ;(Formula 3)

[0091] In formulas 2 and 3 above, E E(t) represents the estimated information exposure parameter of the target object under the i-th candidate stimulus duration; E(t) represents the estimated information exposure parameter of the target object under the reference stimulus duration.

[0092] The click variation parameter reflects the difference between the estimated information click parameter under the candidate stimulus duration and the estimated information click parameter under the reference stimulus duration. This click variation parameter can be either an absolute difference or a relative difference between the estimated information click parameter under the candidate stimulus duration and the estimated information click parameter under the reference stimulus duration.

[0093] That is, click to change parameters It can be calculated using the following formula 4:

[0094] ;(Formula 4)

[0095] Alternatively, it can be calculated using the following formula 5:

[0096] ;(Formula 5)

[0097] In formulas 4 and 5 above, C( C(t) represents the estimated click parameters of the target object under the i-th candidate stimulus duration; C(t) represents the estimated click parameters of the target object under the reference stimulus duration.

[0098] The process of weighting the exposure change parameters and click change parameters for each candidate incentive duration can be described by the following formula:

[0099] ;(Formula 6)

[0100] In a specific embodiment, the smallest candidate stimulus duration among at least two candidate stimulus durations can be used as the reference stimulus duration. In other embodiments, the largest candidate stimulus duration among at least two candidate stimulus durations can also be used as the reference stimulus duration; of course, the reference stimulus duration can also be other candidate stimulus durations among at least two candidate stimulus durations, and no specific limitation is made here.

[0101] In some embodiments, the first weighting coefficient corresponding to the estimated information exposure parameter can be the same for different objects; similarly, the second weighting coefficient corresponding to the estimated information click parameter can also be the same for different objects. In this case, a coefficient can be pre-specified for the estimated information exposure parameter as the first weighting coefficient, and a coefficient can be pre-specified for the estimated information click parameter as the second weighting coefficient.

[0102] In other embodiments, the first weighting coefficient corresponding to the estimated information exposure parameter may be different for different objects, and / or the second weighting coefficient corresponding to the estimated information click parameter may be different for different objects.

[0103] In other embodiments, the first weighting coefficient corresponding to the estimated information exposure parameter may be the same for different objects, for example, it may be 1. For the target object, the second weighting coefficient corresponding to the estimated information click parameter may be determined based on the object hierarchy to which the target object belongs.

[0104] In some embodiments, the step of obtaining the second weighting coefficient corresponding to the predicted information click parameter includes: determining the target object layer to which the target object belongs; determining the weighting coefficient corresponding to the target object layer based on the mapping relationship between the object layer and the weighting coefficient; and using the weighting coefficient corresponding to the target object layer as the second weighting coefficient.

[0105] In some embodiments, objects can be stratified based on their revenue per thousand impressions (RPM) for each object's targeted information. A range of RPM for each object stratum is defined, along with a weighting coefficient. Thus, after determining the RPM for a target object, the target object stratum to which the target object belongs can be determined, and consequently, a second weighting coefficient for the target object can be determined. The RPM for the target object can be calculated based on the number of targeted information messages exposed to the target object in a historical period and the revenue generated from those exposed messages.

[0106] In other embodiments, objects can be stratified based on their click revenue from ad placements over a historical period. Specifically, a click revenue range is defined for each object stratum, and a weighting coefficient is set for each object stratum. Thus, after determining the click revenue for a target object, the target object stratum to which the target object belongs can be determined, and consequently, a second weighting coefficient for the target object can be determined. The click revenue for the target object can be calculated based on the number of ad placements clicked by the target object and the cost-per-click of the clicked ad placements over a historical period.

[0107] It is understandable that for each object, the target excitation duration corresponding to the target period can be determined by following the process of steps 210-230 above.

[0108] In some embodiments, after determining the target incentive duration corresponding to the target object in the target period, the object identifier of the target object is associated with and stored with the corresponding target incentive duration. When the delivery condition is met, the target incentive duration corresponding to the target object is sent to the client where the target object is located, so that the client can deliver virtual resources to the target object according to the corresponding target incentive duration. The delivery condition can be a specified time, which can be set according to actual needs. For example, the specified time can be the start time of the target period, or the delivery condition can be the receipt of the first delivery information request from the target object in the target period. Of course, the delivery condition is not limited to the examples above.

[0109] Interaction duration refers to the length of time an object remains on the page displaying the delivered information after clicking on it. Specifically, if the delivered information is a video, after an object clicks on the video, the video begins playing; in this case, interaction duration can be understood as the playback duration of the delivered information. If the delivered information is text or an image, interaction duration can be understood as the browsing duration of the delivered information. For example, if the delivered information is a blog post, after an object clicks on the blog post, it will enter the blog post's details page. The length of time the object spends on the blog post's details page is the total browsing duration of the blog post.

[0110] In a specific embodiment, the virtual resources may be one or more of the following: lottery opportunities, coupons, feature unlocks, account upgrades, etc., without being specifically limited here.

[0111] In this application, historical data on the target audience's response to the delivered information, along with the target audience's historical incentive duration and characteristic information, are used to predict the estimated information exposure parameters and estimated information click parameters for the target audience under each candidate incentive duration. Since these parameters reflect the target audience's level of attention and interest in the delivered information at each candidate incentive duration, the inventors recognized that setting the incentive duration too long could negatively impact user experience and reduce the reach of the delivered information. Furthermore, different audiences exhibit varying levels of attention and interest in the delivered information even under the same candidate incentive duration. Therefore, this application utilizes these two parameters—the estimated information exposure parameters and the estimated information click parameters—to specifically determine the target incentive duration for the target audience in future target periods. This approach improves the reach of the delivered information and ensures a better user experience by tailoring the target incentive duration to the target audience.

[0112] In some embodiments of this application, the target historical data includes information identifiers for each delivery message exposed to the target object during a historical period; such as... Figure 3 As shown, the steps to determine the object hierarchy to which the target object belongs include:

[0113] Step 310: Based on the target's historical data, count the number of exposures corresponding to each delivery message to the target audience in the historical period.

[0114] Step 320: Based on the information identifiers included in the target historical data, obtain the revenue parameters corresponding to each delivery information exposed to the target audience in the historical period.

[0115] Step 330: Calculate the target revenue parameter for the target object in the historical period based on the number of exposures corresponding to each delivery information targeting the target object in the historical period and the revenue parameters corresponding to each delivery information exposed to the target object in the historical period.

[0116] Step 340: Determine the object stratification corresponding to the target total revenue parameter as the target object stratification to which the target object belongs.

[0117] The revenue parameter for each ad placement can be the revenue earned from placing that ad. It's understandable that the revenue parameter for each ad placement is related to the corresponding billing method.

[0118] In some embodiments, the target revenue parameter corresponding to the target object can be the revenue per thousand impressions (CPM) or the average revenue per impression.

[0119] To calculate the target revenue parameter for the target object, the revenue parameter obtained from each delivery message to the target object in the historical period is divided by the number of times each delivery message is exposed to the target object. This yields the exposure revenue of the target object for each delivery message. Then, the exposure revenue of the target object for all delivery messages is combined to determine the revenue per thousand exposures or the average exposure revenue for the target object.

[0120] In some embodiments, the revenue parameters corresponding to all the delivery information delivered to the target object in the historical period can be added together to obtain the total revenue parameter, and the exposure counts of each delivery information delivered to the target object in the historical period can be added together to obtain the total exposure count. Then, the total revenue parameter can be divided by the total exposure count to calculate the revenue per thousand exposures or the average exposure revenue for the target object.

[0121] Since the target revenue parameter range corresponding to each object layer is preset, after determining the target revenue parameter corresponding to the target object, the target revenue parameter range of the target revenue parameter corresponding to the target object is determined, and the object layer corresponding to the target revenue parameter range of the target total revenue parameter is determined as the object layer to which the target object belongs.

[0122] It is understandable that, since the information exposed to the target object may be different in each period, and the amount of information exposed to the target object may also be different, the target revenue parameters corresponding to the target object will be different for each historical period, and the object layer to which the target object belongs will also change accordingly. Therefore, in this embodiment, it is necessary to redetermine the object layer to which the target object belongs on a periodic basis.

[0123] In some embodiments of this application, such as Figure 4 As shown, step 220 includes:

[0124] Step 410: Combine each candidate stimulus duration with the target historical data, historical stimulus duration and feature information of the target object to obtain at least two sets of input information.

[0125] It is understandable that the number of input information sets is equal to the number of candidate stimulus durations.

[0126] In some embodiments, behavioral features can be extracted from historical target data, and then the behavioral features, historical incentive duration, and feature information of the target object can be combined with each candidate incentive duration to obtain the corresponding input information. The extracted behavioral features may include the information identifier of the clicked information, the exposure tag (exposed or not exposed), the interaction duration of the information, the application identifier of the application that exposed the information, and the location identifier of the exposed position of the information.

[0127] Step 420: The exposure prediction model predicts the exposure parameters based on each set of input information and outputs the predicted exposure parameters of the target object in the future target period under each candidate excitation duration.

[0128] Step 430: The click prediction model predicts the click parameters for each set of input information and outputs the predicted click parameters of the target object in the future target period under each candidate excitation duration.

[0129] In this embodiment, artificial intelligence (AI) technology is used to predict information exposure parameters and information click parameters. Artificial intelligence (AI) is the theory, methods, technology, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. AI also studies the design principles and implementation methods of various intelligent machines, enabling them to possess perception, reasoning, and decision-making capabilities.

[0130] Artificial intelligence (AI) is a comprehensive discipline encompassing a wide range of fields, including both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies primarily include computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0131] Understandably, to ensure the accuracy of the exposure prediction model's predictions of the target information's exposure parameters and the click prediction model's predictions of the target information's click parameters, both models need to be pre-trained. This allows the exposure prediction model to learn the relationship between historical target data, historical stimulus durations, target object features, candidate stimulus durations, and the target object's information exposure parameters in the next cycle, thus enabling it to predict information exposure parameters. Similarly, through training, the click prediction model learns the relationship between historical target data, historical stimulus durations, target object features, candidate stimulus durations, and the target object's information click parameters in the next cycle, thereby enabling it to predict information click parameters.

[0132] The training process for the exposure prediction model and the click prediction model may include the following steps: acquiring training data, which includes multiple training samples. The training samples include at least two candidate incentive durations corresponding to the sample object in the first historical period, the first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, the feature information of the sample object, and the sample incentive duration corresponding to the sample object in the second historical period; wherein, the second historical period is later than the first historical period in terms of time sequence; and training the exposure prediction model and the click prediction model in reverse based on the at least two candidate incentive durations corresponding to the sample object in the first historical period, the first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, the feature information of the sample object, and the sample incentive duration corresponding to the sample object in the second historical period, until the training termination condition is met.

[0133] Specifically, during the training process, each candidate stimulus duration corresponding to the sample object in the first historical period is combined with the first historical data and the feature information of the sample user corresponding to that candidate stimulus duration in the first historical period to obtain at least two sets of sample input information.

[0134] Then, the sample input information of each group is input into the exposure prediction model, which predicts the sample prediction information exposure parameters of the sample object in the second historical period under the candidate excitation duration corresponding to the sample input information. By repeating this process, the sample prediction information exposure parameters of the sample object under each candidate excitation duration in the second historical period can be obtained.

[0135] Furthermore, each set of sample input information is input into the click prediction model, which then predicts the sample prediction information click parameters of the sample object in the second historical period under the candidate excitation duration corresponding to the sample input information. By repeating this process, the sample prediction information click parameters of the sample object under each candidate excitation duration in the second historical period can be obtained.

[0136] Subsequently, based on the sample object's estimated exposure parameters and estimated click parameters for each candidate incentive duration in the second historical period, the target incentive duration for the sample object in the second historical period is determined from at least two candidate incentive durations; whereby the target incentive duration refers to the determined target incentive duration for the sample object in the second historical period. The specific process for determining the target incentive duration is described above and will not be repeated here.

[0137] Subsequently, based on the sample target excitation duration and sample excitation duration corresponding to the sample objects in the second historical period, and the set loss function, the corresponding loss value is calculated. The parameters of the exposure prediction model and the click prediction model are then adjusted in reverse based on the calculated loss value. This allows the adjusted exposure prediction model and click prediction model to re-predict information exposure parameters and information click parameters based on the sample input information. This process is repeated until the training termination condition is met. The training termination condition can be that the number of iterations of the model (exposure prediction model, click prediction model) reaches a set number, or that the loss function converges.

[0138] The loss function can be cross-entropy loss function, absolute value loss function, squared loss function, log loss function, etc., and no specific limitation is made here.

[0139] In one specific embodiment, the set loss function can be the Huber loss function. The Huber loss function is a loss function that combines the MAE (Mean Absolute Error) loss function with the MSE (Mean Squared Error) loss function. Specifically, the Huber loss function is expressed as:

[0140] ;(Formula 7)

[0141] Where y is the sample excitation duration corresponding to the sample object in the second historical period; f(x) is the sample target excitation duration corresponding to the sample object in the second historical period; These are customizable hyperparameters. It can be seen that... For prediction bias, when the prediction bias is no greater than The Huber loss function uses the MSE function; when the prediction bias is greater than... The Huberloss function uses the MAE function.

[0142] Because the MAE function has a large gradient during training, it can cause the minimum value to be missed during gradient descent training. For MSE, the gradient gradually decreases as the prediction bias approaches its minimum, thus improving accuracy. Compared to MAE, the Huber loss function decreases the gradient around its minimum; compared to MSE, the Huber loss function is more robust. Therefore, the Huber loss function combines the advantages of both MAE and MSE functions, thereby improving the training performance of both exposure and click prediction models.

[0143] This application also provides a method for distributing virtual resources. Figure 5 This is a flowchart illustrating a method for distributing virtual resources according to an embodiment of this application. This method can be executed by a terminal, such as a smartphone, tablet, laptop, desktop computer, smart TV, or other electronic device with integrated display functions and processing capabilities. Figure 5 As shown, the method includes:

[0144] Step 510: Display delivery information.

[0145] In some embodiments, after a client in the terminal detects a delivery information request operation triggered by an object, the client sends a delivery information request to the server. The server responds to the delivery information request by returning the corresponding delivery information to the client and displaying it on the client. Delivery information request operations include, for example, page refresh operations, entering the client, playing video operations, triggering specified controls, etc., and are not specifically limited here.

[0146] Step 520: In response to the click operation triggered by the delivery information, count the duration of the interaction between the object and the delivery information.

[0147] In practice, if the information is presented as a video, the interaction duration can be the video playback duration. In this case, the click action triggered by the object on the information can be a video playback action. If the information is presented as text or an image, the interaction duration can be the time spent on the page containing the information. In this case, the timer can start when a click action is detected and stop when the object leaves the page containing the information. The duration from the start of the timer to the stop of the timer is the interaction duration of the object on the information.

[0148] Step 530: If the interaction duration reaches the incentive duration corresponding to the object in the current period, then virtual resources are issued to the object. The incentive duration corresponding to the object in the current period is determined according to the incentive duration determination method shown in any of the above embodiments.

[0149] This allows for the distribution of virtual resources to target audiences based on the specific incentive duration determined for each target audience. This can improve the reach of information delivery and ensure a better user experience for the target audience.

[0150] In some embodiments, the method further includes uploading the exposure records of the delivery information on the client where the object resides, the click records of the object in response to the delivery information, and the total interaction duration of the object in response to the delivery information to the server, so that the server can predict the incentive duration corresponding to the object in the next period. This enables the periodic updating of the incentive duration corresponding to each object, ensuring the matching degree and timeliness of the incentive duration determined for each object.

[0151] The method of this application will now be described in detail with reference to a specific application scenario. In this embodiment, the information delivered is a video advertisement, and the target is the user corresponding to the application. In this embodiment, since virtual resources are distributed to the corresponding user when the playback duration of the video advertisement reaches the corresponding incentive duration, the video advertisement can be called an incentivized video advertisement. Incentivized video advertisement is a form of advertisement in which users actively expose advertisements.

[0152] For rewarded video ads, the set reward duration directly affects the user experience. Understandably, if the reward duration is too long, it can easily cause greater interference to users, affecting their user experience. Moreover, it may also affect whether users will actively request to see the next rewarded video ad, thus affecting the exposure of rewarded video ads on the platform. Of course, it may also affect the platform's advertising revenue.

[0153] In practice, experiments were conducted to investigate the impact of different incentive durations on the decrease in ad impressions and the increase in platform revenue. The specific curves showing the changes in ad impression decrease and revenue increase with incentive duration are as follows: Figure 6As shown.

[0154] In the specific experiment, for all users, the incentive duration was uniformly configured to 20 seconds, 25 seconds, and 30 seconds respectively. The exposure change rate, click change rate, and revenue change rate were statistically analyzed compared to the uniform incentive duration of 15 seconds. The results are shown in Table 2 below (in Table 2, CTR stands for Click-Through-Rate).

[0155] Table 2

[0156]

[0157] comprehensive Figure 6 As can be seen from Table 2 above, the longer the incentive duration, the higher the revenue increase and the greater the exposure decrease. Moreover, there is a diminishing marginal effect in the revenue increase, but no diminishing marginal effect in the exposure decrease. In other words, although a longer incentive duration contributes to the revenue increase, it is more detrimental to the user experience. Therefore, the method of this application can be used to, on the one hand, determine the incentive duration for users in a targeted manner to improve the reach of video ads, and on the other hand, to strike a balance between user experience and revenue.

[0158] Specifically, in this embodiment, considering that user activity behavior on most traffic exhibits periodic fluctuations on a weekly basis, and also to avoid the incentive duration changing too frequently, a week (7 days) is used as the change cycle for the incentive duration and the cumulative cycle for object behavior data.

[0159] For users, the number of times a video ad is displayed can reflect their user experience with the application to some extent, while the number of clicks on a video ad can reflect their level of interest in the ad to some extent. Therefore, in this embodiment, the number of clicks is used as the information click parameter, and the number of displays is used as the information display parameter.

[0160] Considering that in practice, video ad impressions far exceed clicks, and the numerical difference between the two is significant, to unify the units of measurement for impressions and clicks, we calculate the exposure gain ratio (click gain ratio) relative to the impressions (clicks) at the reference incentive duration for each candidate incentive duration. Here, the exposure gain ratio can be considered as the exposure change parameter mentioned above, and the click gain ratio can be considered as the click change parameter mentioned above.

[0161] In this embodiment, the candidate excitation durations are set to 15 seconds, 20 seconds, 25 seconds, and 30 seconds. Using 15 seconds as the reference excitation duration, the exposure gain rate is:

[0162] Exposure gain ratio = (Exposure at T seconds - Exposure at 15 seconds) / Exposure at 15 seconds; (Formula 8)

[0163] Click gain rate:

[0164] Click-through rate = (Number of clicks in T seconds - Number of clicks in 15 seconds) / Number of clicks in 15 seconds; (Formula 9)

[0165] In formulas 8 and 9 above, the value of T is 15, 20, 25 and 30, corresponding to the candidate excitation durations mentioned above.

[0166] Furthermore, to facilitate the selection of candidate excitation durations as the target excitation duration by comprehensively considering exposure gain rate and click gain rate, the exposure gain rate and click gain rate under each candidate excitation duration are weighted, and the target excitation duration is determined based on the weighted results of each candidate excitation duration. The process of weighting the exposure gain rate and click gain rate under each candidate excitation duration can be described by the following formula 10:

[0167] ;(Formula 10)

[0168] Wherein, P(T) represents the weighted result obtained by weighting the exposure gain rate and the click gain rate under the candidate excitation duration of T seconds; it can be understood that since the statistics are performed on a weekly basis, the click gain rate in Formula 10 above can be understood as the cumulative click gain rate of a week, and similarly, the exposure gain rate in Formula 10 above can be understood as the cumulative exposure gain rate of a week.

[0169] As can be seen from Formula 10 above, in this embodiment, the first weighting coefficient is 1, and the second weighting coefficient is... .

[0170] Specifically in this embodiment, This can be determined based on the user segment to which the user belongs. In this embodiment, user segmentation is based on the user's corresponding CPM (Cost Per Mille). Specifically, the user group can be divided into four user segments based on CPM, and a corresponding value can be set for each user segment. In a specific implementation, users with high CPM can be assigned a corresponding lower CPM. Conversely, users with low CPM correspond to higher... .

[0171] After obtaining the click gain rate and exposure gain rate of the user under each candidate incentive duration, the weighting process is performed according to Formula 10 to obtain the weighted result of the user under each candidate incentive duration. The candidate incentive duration with the largest corresponding weighted result is determined as the target incentive duration of the user in the target period.

[0172] In this example, an exposure prediction model is used to predict the amount of time a user is exposed to the ad message at each incentive duration, and a click prediction model is used to predict the amount of time a user clicks on the ad message at each incentive duration.

[0173] In this embodiment, the XGBoost model, which has low tuning costs and simple engineering implementation, is selected as the exposure prediction model and click prediction model. This XGBoost model is a regression model. Considering the distribution characteristics of user clicks and exposures, the majority of samples have low values ​​(close to 0), while some samples have values ​​in the hundreds or thousands. The Huber Loss function is used as the loss function. The mean squared error (MSE) is calculated for low-value samples, and the mean absolute error (MAE) is calculated for high-value samples, thus ensuring that the model has a good fit to most user samples.

[0174] To ensure the accuracy of the exposure prediction model and the click prediction model, they need to be trained first. Figure 7A This is a schematic diagram illustrating the training of an exposure prediction model and a click prediction model based on a specific implementation. For example... Figure 7A As shown, after selecting the sample user group, the user group data of each sample user in the sample user group in week T-1 and the user group data of each sample user in week T are obtained. Features are extracted from the user group data of each sample user in week T-1 and from the user group data of each sample user in week T.

[0175] The features extracted from the user group data of each sample user in week T-1 include: at least two candidate incentive durations corresponding to the sample user in week T-1, the application identifier of the application that exposed the video ad to the sample user in week T-1 (this application identifier can be regarded as traffic scenario feature), the interactive behavior characteristics of the sample user towards the video ad in week T-1 (such as whether the video ad was exposed, the location of the exposed video ad, whether the video ad was clicked, the playback duration of the video ad, etc.), and the feature information of the sample user in week T-1 (such as gender, age, etc.).

[0176] Features extracted from the user group data of each sample user in week T include the sample incentive duration corresponding to the sample user in week T.

[0177] Based on this, the extracted sample users will have at least two candidate incentive durations corresponding to week T-1.

[0178] The training sample consists of the traffic scenario characteristics of the sample users in week T-1, the interaction behavior characteristics of the sample users in week T-1, the object characteristics of the sample users in week T-1, and the sample incentive duration of the sample users in week T. Based on this training sample, the exposure prediction model is trained with the exposure volume of the sample users in week T as the training objective; and based on this training sample, the click volume of the sample users in week T is trained with the click volume of the sample users in week T as the training objective. The specific training process is described above and will not be repeated here.

[0179] After model training, the following data can be combined in the same way: week T object features, week T traffic scenario features, user interaction behavior features in week T, user incentive duration in week T, and the candidate incentive durations (15 / 20 / 25 / 30 seconds) to be estimated in week T+1, to obtain the predicted data. Then, the predicted data is input into the exposure prediction model to predict exposure, and into the click prediction model to predict clicks. Finally, based on the user's exposure and clicks at each candidate incentive duration, the target incentive duration for the user in week T+1 is determined.

[0180] Figure 7B This is a flowchart illustrating the periodic determination of excitation duration according to a specific embodiment of this application. It is based on user data and estimated data used to indicate candidate excitation durations (specifically, as shown in the image). Figure 7B As shown, the candidate excitation durations include 15 seconds, 20 seconds, 25 seconds, and 30 seconds, and the excitation duration is determined periodically according to the following process:

[0181] Step 721: Estimate the number of clicks for users under each candidate incentive duration in the following week. Specifically, the click prediction model estimates the number of clicks based on the user data sample and candidate incentive durations mentioned above. The user data sample includes historical data of users this week (including records of user interaction behavior with each incentive video ad this week, the corresponding historical incentive duration for the user this week, and user characteristic information). Figure 7B As shown, the candidate stimulus durations include 15s, 20s, 25s, and 30s.

[0182] Step 722: Estimate user exposure for the next week under each candidate incentive duration. Specifically, the exposure prediction model estimates exposure based on the user data sample and candidate incentive durations mentioned above.

[0183] Step 723: Determine the user's target incentive duration for the following week. The specific process for determining the target incentive duration is described above and will not be repeated here.

[0184] Step 724: Store the target incentive duration. Specifically, it can be stored in the storage module in the form of key-value pairs of <User ID: Target Incentive Duration>.

[0185] Step 725: Receive a video ad request. This video ad request is initiated based on an ad request action triggered by the user on the client side.

[0186] Step 726: Distribute the target incentive duration corresponding to the user. Specifically, in response to the received video ad request, based on the user identifier carried in the video ad request, the server retrieves the target incentive duration corresponding to the user from the storage module and sends the retrieved target incentive duration to the user's client. In other words, in response to the video ad request, the server sends a video ad to the client.

[0187] Step 727: Receive the interaction behavior records (logs) of the video advertisement reported by the client. The interaction behavior records of the video advertisement may include exposure records for video advertisement exposure, click behavior records generated based on user click behavior of video advertisement, playback duration records of video advertisement, etc. The reported interaction behavior records can be used as new user data samples to predict the target incentive duration for users in the next cycle.

[0188] In practice, experiments were conducted on the method of this embodiment, and the experimental results are shown in Table 3 below. The data in Table 3 represent the rate of change relative to a reference incentive duration of 15 seconds. In this experiment, candidate incentive durations were 15, 20, 25, and 30 seconds. According to the scheme of this embodiment, the target incentive duration for each user was determined, and virtual resources were distributed to users according to this target incentive duration. Subsequently, changes in exposure, click-through rate, and advertising revenue were statistically analyzed, and the statistical results are shown in Table 3 below (in Table 3, CTR (Click-Through-Rate) represents the click-through rate).

[0189] Table 3

[0190]

[0191] Comparing Tables 2 and 3, it can be seen that by adopting the solution of this application, the click volume and exposure volume of each user in the next period are predicted in a targeted manner, and the target incentive duration for users in the next period is determined based on the predicted click volume and exposure volume, the drop in exposure volume is greatly reduced. Thus, it is shown that the solution of this application can effectively guarantee user experience.

[0192] The following describes an apparatus embodiment of this application, which can be used to perform the methods described in the above embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the method embodiments described in the above embodiments of this application.

[0193] Figure 8 This is a block diagram of an excitation duration determination device according to an embodiment of this application, such as... Figure 8 As shown, the device for determining the incentive duration includes: an acquisition module 810, used to acquire historical data of the target object in response to the information being delivered during a historical period, the historical incentive duration corresponding to the target object in the historical period, and the characteristic information of the target object; a prediction module 820, used to predict the estimated information exposure parameters and estimated information click parameters corresponding to the target object in the future target period under each candidate incentive duration based on the historical data, historical incentive duration, and characteristic information of the target object; and a target incentive duration determination module 830, used to determine the target incentive duration from at least two candidate incentive durations based on the estimated information exposure parameters and information click parameters corresponding to the target object in the future target period under each candidate incentive duration, wherein if the interaction duration of the target object in response to the information being delivered during the target period reaches the target incentive duration, virtual resources are distributed to the target object.

[0194] In some embodiments, the target incentive duration determination module 830 includes: a first weighting coefficient acquisition unit, configured to acquire a first weighting coefficient corresponding to the estimated information exposure parameter; a second weighting coefficient acquisition unit, configured to acquire a second weighting coefficient corresponding to the estimated information click parameter; a first weighting processing unit, configured to perform weighting processing on the estimated information exposure parameter and the estimated information click parameter corresponding to the target object in the future target period under each candidate incentive duration, based on the first weighting coefficient corresponding to the estimated information exposure parameter and the second weighting coefficient corresponding to the estimated information click parameter, to obtain a weighted result corresponding to each candidate incentive duration; and a target incentive duration determination unit, configured to determine the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

[0195] In other embodiments, the target excitation duration determination module 830 includes: an exposure change parameter calculation unit, configured to calculate the exposure change parameter of the target object relative to a reference excitation duration under each candidate excitation duration based on the estimated information exposure parameter corresponding to the target object under each candidate excitation duration, wherein the reference excitation duration is one of at least two candidate excitation durations; a click change parameter calculation unit, configured to calculate the click change parameter of the target object relative to a reference excitation duration under each candidate excitation duration based on the estimated information click parameter corresponding to the target object under each candidate excitation duration; and a first weighting coefficient acquisition unit. The system comprises: a first weighting coefficient acquisition unit for obtaining the first weighting coefficient corresponding to the estimated information exposure parameter; a second weighting coefficient acquisition unit for obtaining the second weighting coefficient corresponding to the estimated information click parameter; a second weighting processing unit for weighting the exposure change parameter and click change parameter corresponding to each candidate incentive duration based on the first weighting coefficient and the second weighting coefficient corresponding to the estimated information click parameter, to obtain the weighted result corresponding to each candidate incentive duration; and a target incentive duration determination unit for determining the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

[0196] In some embodiments, the second weighting coefficient acquisition unit includes: an object layer determination unit, configured to determine the target object layer to which the target object belongs; and a coefficient determination unit, configured to determine the weighting coefficient corresponding to the target object layer based on the mapping relationship between the object layer and the weighting coefficient, and use the weighting coefficient corresponding to the target object layer as the second weighting coefficient.

[0197] In some embodiments, the target historical data includes information identifiers of each delivery message exposed to the target object in the historical period; in this embodiment, the object layer determination unit includes: an exposure count counting unit, used to count the exposure counts corresponding to each delivery message to the target object in the historical period based on the target historical data; a revenue parameter acquisition unit, used to acquire the revenue parameters corresponding to each delivery message exposed to the target object in the historical period based on the information identifiers included in the target historical data; a target total revenue parameter calculation unit, used to calculate the target revenue parameter corresponding to the target object in the historical period based on the exposure counts corresponding to each delivery message to the target object in the historical period and the revenue parameters corresponding to each delivery message exposed to the target object in the historical period; and a determination unit, used to determine the object layer corresponding to the target total revenue parameter as the target object layer to which the target object belongs.

[0198] In some embodiments, the prediction module 820 includes: a combination unit, configured to combine each candidate excitation duration with target historical data, historical excitation duration, and feature information of the target object to obtain at least two sets of input information; a first prediction unit, configured to use an exposure prediction model to predict exposure parameters based on each set of input information, and output the predicted exposure parameters of the target object in the future target period under each candidate excitation duration; and a second prediction unit, configured to use a click prediction model to predict click parameters based on each set of input information, and output the predicted click parameters of the target object in the future target period under each candidate excitation duration.

[0199] In some embodiments, the device for determining the incentive duration further includes: a training data acquisition module, configured to acquire training data, the training data including multiple training samples, the training samples including at least two candidate incentive durations corresponding to the sample object in a first historical period, first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, feature information of the sample object, and sample incentive duration corresponding to the sample object in a second historical period; wherein, in terms of time sequence, the second historical period is later than the first historical period; the training module is configured to back-train the exposure prediction model and the click prediction model based on at least two candidate incentive durations corresponding to the sample object in the first historical period, first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, object feature information of the sample object, and sample incentive duration corresponding to the sample object in the second historical period, until the training termination condition is reached.

[0200] This application also provides a device for distributing virtual resources. Figure 9 This is a block diagram of a virtual resource distribution device according to an embodiment of this application, such as... Figure 9 As shown, the device includes: a display module 910 for displaying delivery information; a statistics module 920 for counting the interaction duration of the object in response to a click operation triggered by the delivery information; and a distribution module 930 for distributing virtual resources to the object if the interaction duration reaches the incentive duration corresponding to the object in the current period, wherein the incentive duration corresponding to the object in the current period is determined according to the method of any of the above-described methods for determining incentive duration.

[0201] In some embodiments, the virtual resource distribution device further includes: a reporting module, used to upload the exposure record of the distribution information on the client where the object is located, the click record of the object on the distribution information, and the total interaction time record of the object on the distribution information to the server, so as to use the server to predict the incentive time of the object in the next cycle.

[0202] Figure 10A schematic diagram of a computer system suitable for implementing the embodiments of this application is shown. It should be noted that... Figure 10 The computer system 1000 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0203] like Figure 10 As shown, the computer system 1000 includes a processor, which may be a Central Processing Unit (CPU) 1001. The CPU can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 1002 or programs loaded from storage portion 1008 into Random Access Memory (RAM) 1003, such as performing the methods described in the above embodiments. The RAM 1003 also stores various programs and data required for system operation. The CPU 1001, ROM 1002, and RAM 1003 are interconnected via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0204] The following components are connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, etc.; an output section 1007 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. Removable media 1011, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1010 as needed so that computer programs read from them can be installed into storage section 1008 as needed.

[0205] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1009, and / or installed from removable medium 1011. When the computer program is executed by central processing unit (CPU) 1001, it performs various functions defined in the system of this application.

[0206] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0207] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0208] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0209] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium carries computer-readable instructions that, when executed by a processor, implement the methods in any of the above embodiments.

[0210] According to one aspect of this application, an electronic device is also provided, comprising: a processor; and a memory storing computer-readable instructions that, when executed by the processor, implement the methods of any of the above embodiments.

[0211] According to one aspect of the embodiments of this application, 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 the processor executes the computer instructions, causing the computer device to perform the methods of any of the above embodiments.

[0212] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0213] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the methods according to the embodiments of this application.

[0214] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0215] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for determining the excitation duration, characterized in that, include: Acquire the target object's historical data on the information delivery in a historical period, the historical incentive duration corresponding to the target object in the historical period, and the characteristic information of the target object; The target historical data includes interactive behavior records generated based on the interactive behaviors triggered by the target object in response to the delivered information during the historical period; the feature information includes at least one of gender, age, location, marital status, interest tags, account level, and membership status. Based on the target's historical data, the historical incentive duration, and the target object's feature information, predict the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration; Based on the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration, a target incentive duration is determined from at least two candidate incentive durations. If the target object's interaction duration with the delivered information reaches the target incentive duration in the target period, virtual resources are issued to the target object. The virtual resources include at least one of the following: lottery opportunities, coupons, feature unlocks, and account upgrades.

2. The method according to claim 1, characterized in that, The step of determining the target incentive duration from at least two candidate incentive durations based on the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration includes: Obtain a first weighting coefficient corresponding to the estimated information exposure parameter, and obtain a second weighting coefficient corresponding to the estimated information click parameter; Based on the first weighting coefficient corresponding to the estimated information exposure parameter and the second weighting coefficient corresponding to the estimated information click parameter, the estimated information exposure parameter and the estimated information click parameter of the target object in the future target period under each candidate incentive duration are weighted to obtain the weighted result corresponding to each candidate incentive duration. The candidate incentive duration with the largest corresponding weighted result is determined as the target incentive duration.

3. The method according to claim 1, characterized in that, The step of determining the target incentive duration from at least two candidate incentive durations based on the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration includes: Based on the estimated information exposure parameters of the target object corresponding to each candidate incentive duration, calculate the exposure change parameters of the target object at each candidate incentive duration relative to the reference incentive duration, wherein the reference incentive duration is one of at least two candidate incentive durations. Based on the estimated click parameters of the target object corresponding to each candidate incentive duration, calculate the click change parameters of the target object at each candidate incentive duration relative to the reference incentive duration. Obtain a first weighting coefficient corresponding to the estimated information exposure parameter, and obtain a second weighting coefficient corresponding to the estimated information click parameter; Based on the first weighting coefficient corresponding to the estimated information exposure parameter and the second weighting coefficient corresponding to the estimated information click parameter, the exposure change parameter and click change parameter corresponding to each candidate incentive duration are weighted to obtain the weighted result corresponding to each candidate incentive duration. The candidate incentive duration with the largest corresponding weighted result is determined as the target incentive duration.

4. The method according to claim 2 or 3, characterized in that, The step of obtaining the second weighting coefficient corresponding to the predicted information click parameter includes: Determine the target object hierarchy to which the target object belongs; Based on the mapping relationship between object layering and weighting coefficients, the weighting coefficients corresponding to the target object layering are determined, and the weighting coefficients corresponding to the target object layering are used as the second weighting coefficients.

5. The method according to claim 4, characterized in that, The target historical data also includes information identifiers for each delivery information exposed to the target object during the historical period; Determining the target object hierarchy to which the target object belongs includes: Based on the target's historical data, the number of exposures corresponding to each delivery message to the target object during the historical period is counted. Based on the information identifiers included in the target historical data, obtain the revenue parameters corresponding to each delivery information exposed to the target object in the historical period; Based on the number of exposures corresponding to each delivery message targeting the target object in the historical period and the revenue parameters corresponding to each delivery message exposed to the target object in the historical period, the target revenue parameter corresponding to the target object in the historical period is calculated. The object stratification corresponding to the target total revenue parameter is determined as the target object stratification to which the target object belongs.

6. The method according to claim 1, characterized in that, The step of predicting the estimated information exposure parameters and estimated information click parameters of the target object in future target periods under each candidate incentive duration, based on the target's historical data, the historical incentive duration, and the target object's feature information, includes: Each candidate stimulus duration is combined with the target historical data, the historical stimulus duration, and the feature information of the target object to obtain at least two sets of input information; The exposure prediction model predicts the exposure parameters based on each set of input information and outputs the predicted exposure parameters of the target object in the future target period under each candidate excitation duration. The click prediction model predicts click parameters based on each set of input information and outputs the predicted click parameters of the target object in the future target period under each candidate excitation duration.

7. The method according to claim 6, characterized in that, Before the exposure prediction model predicts exposure parameters for each set of input information and outputs the predicted exposure parameters for the target object in the future target period under each candidate excitation duration, the method further includes: Acquire training data, which includes multiple training samples. The training samples include at least two candidate incentive durations corresponding to the sample object in a first historical period, first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, feature information of the sample object, and sample incentive duration corresponding to the sample object in a second historical period; wherein, in terms of time sequence, the second historical period is later than the first historical period. Based on at least two candidate incentive durations corresponding to the sample object in the first historical period, the first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, the feature information of the sample object, and the sample incentive duration corresponding to the sample object in the second historical period, the exposure prediction model and the click prediction model are trained in reverse until the training termination condition is met.

8. A method for distributing virtual resources, characterized in that, include: Display delivery information; In response to a click operation triggered by the delivery information, the duration of the interaction between the object and the delivery information is statistically analyzed. If the interaction duration reaches the incentive duration corresponding to the object in the current period, then virtual resources are issued to the object, wherein the incentive duration corresponding to the object in the current period is determined according to the method of any one of claims 1-7.

9. The method according to claim 8, characterized in that, The method further includes: The server uploads at least one of the following to the server: the exposure record of the target client, the click record of the target on the target, and the total interaction time of the target on the target. This information is used by the server to predict the incentive duration of the target in the next cycle.

10. A device for determining the excitation duration, characterized in that, include: The acquisition module is used to acquire the target historical data of the target object in the historical period for the delivery information, the historical incentive duration of the target object in the historical period, and the feature information of the target object; The target historical data includes interactive behavior records generated based on the interactive behaviors triggered by the target object in response to the delivered information during the historical period; the feature information includes at least one of gender, age, location, marital status, interest tags, account level, and membership status. The prediction module is used to predict the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration, based on the target historical data, the historical incentive duration, and the feature information of the target object. The target incentive duration determination module is used to determine the target incentive duration from at least two candidate incentive durations based on the estimated information exposure parameters and estimated information click parameters of the target object in the future target period under each candidate incentive duration. If the interaction duration of the target object with the delivered information reaches the target incentive duration in the target period, virtual resources are issued to the target object. The virtual resources include at least one of the following: lottery opportunities, coupons, feature unlocking, and account upgrades.

11. The apparatus according to claim 10, characterized in that, The target excitation duration determination module includes: The first weighting coefficient acquisition unit is used to acquire a first weighting coefficient corresponding to the estimated information exposure parameter and a second weighting coefficient corresponding to the estimated information click parameter. The second weighting coefficient acquisition unit is used to perform weighting processing on the estimated information exposure parameters and the estimated information click parameters corresponding to the target object in the future target period under each candidate incentive duration, based on the first weighting coefficient corresponding to the estimated information exposure parameters and the second weighting coefficient corresponding to the estimated information click parameters, to obtain the weighting result corresponding to each candidate incentive duration. The first weighted processing unit is used to determine the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

12. The apparatus according to claim 10, characterized in that, The target excitation duration determination module includes: An exposure change parameter calculation unit is used to calculate the exposure change parameter of the target object relative to the reference excitation duration based on the estimated information exposure parameter of the target object corresponding to each candidate excitation duration, wherein the reference excitation duration is one of at least two candidate excitation durations. The click change parameter calculation unit is used to calculate the click change parameter of the target object at each of the candidate incentive durations relative to the reference incentive duration, based on the estimated information click parameter of the target object at each candidate incentive duration. The first weighting coefficient acquisition unit is used to acquire a first weighting coefficient corresponding to the estimated information exposure parameter and a second weighting coefficient corresponding to the estimated information click parameter. The second weighting coefficient acquisition unit is used to perform weighting processing on the exposure change parameters and click change parameters corresponding to each candidate incentive duration based on the first weighting coefficient corresponding to the estimated information exposure parameters and the second weighting coefficient corresponding to the estimated information click parameters, so as to obtain the weighted result corresponding to each candidate incentive duration. The target incentive duration determination unit is used to determine the candidate incentive duration with the largest corresponding weighted result as the target incentive duration.

13. The apparatus according to claim 11 or 12, characterized in that, The second weighting coefficient acquisition unit includes: An object layer determination unit is used to determine the target object layer to which the target object belongs; The coefficient determination unit is used to determine the weighting coefficient corresponding to the target object layer based on the mapping relationship between the object layer and the weighting coefficient, and to use the weighting coefficient corresponding to the target object layer as the second weighting coefficient.

14. The apparatus according to claim 13, characterized in that, The target historical data also includes information identifiers for each delivery information exposed to the target object during the historical period; the object layering determination unit includes: An exposure count counting unit is used to count the number of exposures corresponding to each delivery message to the target object in the historical period based on the target's historical data. The revenue parameter acquisition unit is used to acquire the revenue parameters corresponding to each delivery information exposed to the target object in the historical period based on the information identifiers included in the target historical data. The target total revenue parameter calculation unit is used to calculate the target revenue parameter corresponding to the target object in the historical period based on the number of exposures corresponding to each delivery information for the target object in the historical period and the revenue parameters corresponding to each delivery information exposed to the target object in the historical period. The determining unit is used to determine the object layer corresponding to the target total revenue parameter as the target object layer to which the target object belongs.

15. The apparatus according to claim 10, characterized in that, The prediction module includes: The combination unit is used to combine each candidate stimulus duration with the target historical data, the historical stimulus duration, and the feature information of the target object to obtain at least two sets of input information; The first prediction unit is used to predict the exposure parameters by the exposure prediction model based on each set of input information, and output the predicted exposure parameters of the target object in the future target period under each candidate excitation duration. The second prediction unit is used to predict the click parameters based on each set of input information by the click prediction model, and output the predicted click parameters of the target object in the future target period under each candidate excitation duration.

16. The apparatus according to claim 15, characterized in that, The device for determining the excitation duration further includes: The training data acquisition module is used to acquire training data, which includes multiple training samples. The training samples include at least two candidate incentive durations corresponding to the sample object in a first historical period, first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, feature information of the sample object, and sample incentive duration corresponding to the sample object in a second historical period; wherein, the second historical period is later than the first historical period in terms of time sequence. The training module is used to back-train the exposure prediction model and the click prediction model based on at least two candidate incentive durations corresponding to the sample object in the first historical period, the first historical data of the sample object for the delivery information under each candidate incentive duration in the first historical period, the feature information of the sample object, and the sample incentive duration corresponding to the sample object in the second historical period, until the training termination condition is reached.

17. A device for distributing virtual resources, characterized in that, include: The display module is used to display delivery information; The statistics module is used to respond to click operations triggered by the delivery information and to count the interaction duration of the delivery information. The distribution module is used to distribute virtual resources to the object if the interaction duration reaches the incentive duration corresponding to the object in the current period, wherein the incentive duration corresponding to the object in the current period is determined according to the method of any one of claims 1-7.

18. The apparatus according to claim 17, characterized in that, The virtual resource distribution device further includes: The reporting module is used to upload at least one of the following to the server: the exposure record of the advertising information on the client where the object is located, the click record of the object in response to the advertising information, and the total interaction time record of the object in response to the advertising information, so that the server can predict the incentive time of the object in the next cycle.

19. An electronic device, characterized in that, include: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1-9.

20. A computer-readable storage medium storing computer-readable instructions thereon, characterized in that, When the computer-readable instructions are executed by a processor, the method as described in any one of claims 1-9 is implemented.

21. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the method of any one of claims 1-9.

Images