Pop-up display, target probability prediction model training method and device

Abstract

The present disclosure relates to a pop-up display, a target probability prediction model training method and device. The method comprises: after detecting that a target application is started, obtaining content interaction features of a user account in the target application; the content interaction features are features obtained according to interaction information between the user account and content in the target application; inputting the content interaction features into a target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents a probability that the user account closes the target application at a target time; the second prediction probability represents an association probability between the user account and content corresponding to a to-be-recommended pop-up window; and the to-be-recommended pop-up window is displayed according to the first prediction probability and the second prediction probability. The present disclosure can improve the utilization rate of an application program.

Description

Technical Field

[0001] This disclosure relates to the field of Internet technology, and in particular to pop-up display, training methods, devices and storage media for target probability prediction models. Background Technology

[0002] Most in-app pop-up methods are triggered at specific times, making personalization impossible. However, different users, and even the same user at different times, have different needs for pop-ups. Triggering pop-ups at specific times not only leads to inefficiency and fails to meet the platform's requirements for handling pop-up activities, but more importantly, it may interfere with users' normal content access behavior, which is not conducive to increasing users' usage time of the application.

[0003] Therefore, traditional technologies suffer from low application usage rates. Summary of the Invention

[0004] This disclosure provides a pop-up window display method, apparatus, electronic device, and storage medium to at least address the problem of low application usage rates in related technologies. The technical solution of this disclosure is as follows:

[0005] According to a first aspect of the present disclosure, a pop-up window display method is provided, comprising:

[0006] Upon detecting the launch of a target application, the user account's content interaction characteristics within the target application are obtained; these content interaction characteristics are features obtained based on the interaction information between the user account and the content in the target application.

[0007] The content interaction features are input into the target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents the probability that the user account closes the target application at a target time; the second prediction probability represents the probability that the user account is associated with the content corresponding to the pop-up to be recommended.

[0008] The pop-up window to be recommended is displayed based on the first predicted probability and the second predicted probability.

[0009] In one possible implementation, the first predicted probability and the second predicted probability are used to display the pop-up window to be recommended, including:

[0010] If the first predicted probability is greater than a preset first probability threshold and the second predicted probability is greater than a preset second probability threshold, then the step of displaying the pop-up window to be recommended is executed.

[0011] or,

[0012] The first predicted probability and the second predicted probability are weighted and summed to obtain the fused predicted probability; if the fused predicted probability is greater than a preset third probability threshold, then the step of displaying the pop-up window to be recommended is executed.

[0013] In one possible implementation, the method further includes:

[0014] A model configuration retrieval request is sent to the server; the model configuration retrieval request is used to instruct the server to return the model configuration parameters corresponding to the target probability prediction model;

[0015] Receive the model configuration parameters returned by the server;

[0016] The target probability prediction model is constructed based on the model configuration parameters.

[0017] In one possible implementation, if the content corresponding to the pop-up to be recommended is a target virtual resource, after the step of displaying the pop-up to be recommended based on the first prediction probability and the second prediction probability, the method further includes:

[0018] In response to the claim operation of the target virtual resource, the virtual resource is transferred to the virtual resource account corresponding to the user account.

[0019] According to a second aspect of the present disclosure, a training method for the above-described target probability prediction model is provided, the method comprising:

[0020] Acquire training sample data; the training sample data includes a first training log and a second training log; the first training log records the content interaction information of the sample user account in the target application, as well as the event information of the sample user account closing the target application; the second training log records the content interaction information of the sample user account in the target application, as well as the event information of the sample user account interacting with the content corresponding to the pop-up to be recommended;

[0021] The initial probability prediction model is trained using the training sample data to obtain the target probability prediction model.

[0022] In one possible implementation, training the initial probability prediction model using the training sample data to obtain the target probability prediction model includes:

[0023] In the first training log, content interaction sample features are determined based on the content interaction features within a preset time period before the time to be predicted.

[0024] The content interaction sample features are input into the first probability prediction model in the initial probability prediction model to obtain a first prediction output result; the first prediction output result includes the predicted probability that the sample user account closes the target application at the time to be predicted;

[0025] In the first training log, it is determined whether there is an event information of the sample user account closing the target application recorded at the time to be predicted, so as to obtain the first target output result. Based on the difference between the first target output result and the first prediction output result, the first probability prediction model in the initial probability prediction model is trained until the difference is less than a preset difference threshold.

[0026] In one possible implementation, training the initial probability prediction model using the training sample data to obtain the target probability prediction model includes:

[0027] In the second training log, content interaction sample features are determined based on the content interaction features within a preset time period before the time to be predicted.

[0028] The content interaction sample features are input into the second probability prediction model in the initial probability prediction model to obtain the second prediction output result; the second prediction output result includes the association probability between the sample user account and the content corresponding to the pop-up to be recommended;

[0029] In the second training log, it is determined whether there is an event information recording the interaction between the sample user account and the content corresponding to the pop-up to be recommended at the time to be predicted, so as to obtain the second target output result. Based on the difference between the second target output result and the second prediction output result, the second probability prediction model in the initial probability prediction model is trained until the difference is less than the preset difference threshold.

[0030] In one possible implementation, acquiring the training sample data includes:

[0031] Obtain the content interaction log of the sample user account; the content interaction log records the content interaction information of the sample user account in the target application.

[0032] Obtain the application shutdown log of the sample user account; the application shutdown log records event information of the sample user account closing the target application within the preset time period;

[0033] The first training log is obtained by concatenating the content interaction log and the application closing log.

[0034] In one possible implementation, the method further includes:

[0035] Obtain the pop-up interaction log of the sample user account; the pop-up interaction log records event information of the sample user account interacting with the content corresponding to the pop-up to be recommended within the preset time period;

[0036] The second training log is obtained by concatenating the content interaction log and the pop-up interaction log.

[0037] According to a third aspect of the present disclosure, a pop-up window display device is provided, comprising:

[0038] The acquisition unit is configured to acquire, upon detecting the launch of a target application, content interaction features of a user account in the target application; the content interaction features are features obtained based on interaction information between the user account and content in the target application.

[0039] The determining unit is configured to input the content interaction features into a target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents the probability that the user account closes the target application at a target time; the second prediction probability represents the probability that the user account is associated with the content corresponding to the pop-up to be recommended.

[0040] The display unit is configured to display the pop-up window to be recommended based on the first predicted probability and the second predicted probability.

[0041] In one possible implementation, the display unit is specifically configured to execute the step of displaying the pop-up window to be recommended if the first predicted probability is greater than a preset first probability threshold and the second predicted probability is greater than a preset second probability threshold; or, to perform a weighted summation of the first predicted probability and the second predicted probability to obtain a fused predicted probability; and to execute the step of displaying the pop-up window to be recommended if the fused predicted probability is greater than a preset third probability threshold.

[0042] In one possible implementation, the apparatus further includes: a sending module configured to send a model configuration acquisition request to a server; the model configuration acquisition request being used to instruct the server to return model configuration parameters corresponding to the target probability prediction model; a receiving module configured to receive the model configuration parameters returned by the server; and a construction module configured to construct the target probability prediction model based on the model configuration parameters.

[0043] In one possible implementation, the apparatus further includes a transfer module configured to perform a transfer of the virtual resource to a virtual resource account corresponding to the user account in response to a claim operation on the target virtual resource.

[0044] According to a fourth aspect of the present disclosure, a training apparatus for the above-described target probability prediction model is provided, comprising:

[0045] The sampling unit is configured to acquire training sample data; the training sample data includes a first training log and a second training log; the first training log records content interaction information of the sample user account in the target application, as well as event information of the sample user account closing the target application; the second training log records content interaction information of the sample user account in the target application, as well as event information of the sample user account interacting with the content corresponding to the pop-up to be recommended.

[0046] The training unit is configured to train the initial probability prediction model using the training sample data to obtain the target probability prediction model.

[0047] In one possible implementation, the training unit is specifically configured to execute in the first training log, determining content interaction sample features based on content interaction features within a preset time period prior to the time to be predicted; inputting the content interaction sample features into a first probability prediction model in the initial probability prediction model to obtain a first prediction output result; the first prediction output result includes the predicted probability that the sample user account closes the target application at the time to be predicted; in the first training log, determining whether there is an event information recorded at the time to be predicted that the sample user account closes the target application to obtain a first target output result, and training the first probability prediction model in the initial probability prediction model based on the difference between the first target output result and the first prediction output result until the difference is less than a preset difference threshold.

[0048] In one possible implementation, the training unit is specifically configured to execute in the second training log to determine content interaction sample features based on content interaction features within a preset time period before the time to be predicted; input the content interaction sample features into the second probability prediction model in the initial probability prediction model to obtain a second prediction output result; the second prediction output result includes the association probability between the sample user account and the content corresponding to the pop-up to be recommended; in the second training log, determine whether there is event information recording the interaction between the sample user account and the content corresponding to the pop-up to be recommended at the time to be predicted, obtain a second target output result, and train the second probability prediction model in the initial probability prediction model based on the difference between the second target output result and the second prediction output result until the difference is less than a preset difference threshold.

[0049] In one possible implementation, the sampling unit is specifically configured to perform the following actions: obtaining the content interaction log of the sample user account; the content interaction log records the content interaction information of the sample user account in the target application; obtaining the application closure log of the sample user account; the application closure log records the event information of the sample user account closing the target application within the preset time period; and concatenating the content interaction log and the application closure log to obtain the first training log.

[0050] In one possible implementation, the sampling unit is specifically configured to perform the following: obtain the pop-up interaction log of the sample user account; the pop-up interaction log records event information of the sample user account interacting with the content corresponding to the pop-up to be recommended within the preset time period; and concatenate the content interaction log and the pop-up interaction log to obtain the second training log.

[0051] According to a fifth aspect of the present disclosure, an electronic device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement a pop-up display method as described in the first aspect or any possible implementation of the first aspect, and / or a training method as described in the second aspect or any possible implementation of the second aspect.

[0052] According to a sixth aspect of the present disclosure, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the pop-up display method as described in the first aspect or any possible implementation thereof, and / or the training method as described in the second aspect or any possible implementation thereof.

[0053] According to a seventh aspect of the present disclosure, a computer program product is provided, the program product comprising a computer program stored in a readable storage medium, wherein at least one processor of a device reads from the readable storage medium and executes the computer program, causing the device to perform the pop-up display method described in the first aspect or any possible implementation thereof, and / or the training method described in the second aspect or any possible implementation thereof.

[0054] The technical solution provided by the embodiments of this disclosure brings at least the following beneficial effects: After detecting the launch of the target application, content interaction features are obtained by acquiring interaction information between the user account and the content in the target application; the content interaction features are input into the target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents the probability that the user account will close the target application at a target time; the second prediction probability represents the association probability between the user account and the content corresponding to the pop-up to be recommended; and the pop-up to be recommended is displayed according to the first prediction probability and the second prediction probability. In this way, by acquiring content interaction features, the user's interactive behavior in the target application can be captured in real time, and if it is determined based on the content interaction features that the user will exit the target application at a target time, a pop-up to be recommended that matches the user's personalized preferences can be displayed to the user. This can be achieved by using the pop-up to be recommended to guide the user account to continue accessing the content provided by the target application, thereby increasing the time the user account accesses the target application and improving the effective usage rate of the target application.

[0055] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0056] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0057] Figure 1 This is a flowchart illustrating a pop-up window display method according to an exemplary embodiment.

[0058] Figure 2 This is a training method for a behavior prediction model according to an exemplary embodiment.

[0059] Figure 3 This is a flowchart illustrating a step of obtaining training sample data according to an exemplary embodiment.

[0060] Figure 4 This is another method for displaying a pop-up window according to an exemplary embodiment.

[0061] Figure 5 This is a flowchart illustrating a pop-up window display method according to an exemplary embodiment.

[0062] Figure 6 This is a block diagram illustrating a pop-up display device according to an exemplary embodiment.

[0063] Figure 7 This is a block diagram illustrating a training apparatus for a behavior prediction model according to an exemplary embodiment.

[0064] Figure 8 This is a block diagram illustrating an electronic device according to an exemplary embodiment.

[0065] Figure 9 This is a block diagram illustrating a server according to an exemplary embodiment. Detailed Implementation

[0066] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0067] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0068] It should also be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0069] Figure 1 This is a flowchart illustrating a pop-up window display method according to an exemplary embodiment, such as... Figure 1 As shown, this pop-up display method is used in electronic devices and includes the following steps.

[0070] In step S110, after detecting that the target application has started, the content interaction features corresponding to the target application are obtained.

[0071] Among them, the content interaction feature is a feature obtained based on the interaction information between the user account and the content in the target application.

[0072] In practical applications, content interaction features can be features constructed based on the application operation sequence of a user account when accessing content in the target application, or features constructed based on the content interaction behavior of a user account in the target application.

[0073] Taking a short video application as an example, content interaction behavior can refer to the user's account's behavior on the short video application, such as watching videos, liking videos, and following authors.

[0074] In practical implementation, after a user account launches a target application through an electronic terminal, the electronic device can collect the user account's content access behavior within the target application at a preset frequency. Then, the electronic device performs feature extraction processing on the collected content access behavior (e.g., application operation sequences, content interaction behaviors, etc.) to obtain the user account's corresponding content interaction features in the target application, and caches these features on the electronic device. In practical applications, due to resource limitations of the electronic device, the feature caching of content interaction features requires a feature eviction mechanism to ensure that the feature storage space does not explode. This disclosure can cache content interaction features within a preset time period before the target time, i.e., the user's most recent N consumption behaviors, in the electronic device, and permanently protect content interaction features whose importance exceeds a preset threshold in the cache, ensuring that content interaction features with importance exceeding the preset threshold will not be evicted.

[0075] In step S120, the content interaction features are input into the target probability prediction model to obtain the first prediction probability and the second prediction probability.

[0076] The first predicted probability represents the probability that a user account will close the target application at the target time.

[0077] The second predicted probability represents the probability of association between a user account and the content corresponding to the pop-up to be recommended. In practical applications, the association probability can be used to represent the probability of a user account's interest in the content corresponding to the pop-up to be recommended.

[0078] Of course, the probability of associating a user account with the content corresponding to the pop-up to be recommended can be used to characterize the probability of interaction between the user account and the pop-up to be recommended, i.e., the pop-up click-through rate.

[0079] The "Pending Recommendation" pop-up is used to guide user accounts to continue accessing content provided by the target application. In practical applications, the "Pending Recommendation" pop-up can be a pop-up for distributing virtual resources (i.e., an electronic red envelope pop-up), or it can be a dialog pop-up that includes an entry point to view the content to be recommended (e.g., a dialog pop-up that includes an entry point to view the live stream of the live stream to be recommended).

[0080] In practice, after acquiring content interaction features, the electronic device can input these features into a target probability prediction model. This model then classifies the content interaction features and, based on the classification results, determines whether the user account will close the target application within a preset time period, and whether the user account will click on a preset recommended pop-up window. The classification results include a first predicted probability and a second predicted probability. When the first predicted probability is greater than a preset probability threshold, the electronic device determines that the user account will close the target application within the preset time period; when the second predicted probability is greater than the preset probability threshold, the electronic device determines that the user account will click on the preset recommended pop-up window.

[0081] In step S130, a pop-up window for recommendations is displayed based on the first predicted probability and the second predicted probability.

[0082] In practice, the electronic device can determine whether it should trigger a push notification (i.e., a pop-up to be recommended) within the target application based on a first predicted probability and a second predicted probability. Specifically, when the electronic device determines that the user account will close the target application within a preset time period and that the user account will click on the preset pop-up to be recommended, the electronic device will trigger the push notification within the target application, i.e., display the pop-up to be recommended, thereby guiding the user account to continue accessing the content provided by the target application.

[0083] In the aforementioned pop-up display method, after detecting the launch of the target application, content interaction features are obtained by acquiring interaction information between the user account and the content in the target application. These content interaction features are then input into a target probability prediction model to obtain a first prediction probability and a second prediction probability. The first prediction probability represents the probability that the user account will close the target application at a target time. The second prediction probability represents the association probability between the user account and the content corresponding to the pop-up to be recommended. Based on the first and second prediction probabilities, the pop-up to be recommended is displayed. Thus, by acquiring content interaction features, the user's interactive behavior in the target application can be captured in real time. If, based on the content interaction features, it is determined that the user will exit the target application at a target time, a pop-up to be recommended that matches the user's personalized preferences is displayed. This allows the user account to continue accessing the content provided by the target application, thereby increasing the time the user account spends accessing the target application and improving the effective usage rate of the target application.

[0084] In an exemplary embodiment, displaying the pop-up window to be recommended using a first predicted probability and a second predicted probability includes: if the first predicted probability is greater than a preset first probability threshold and the second predicted probability is greater than a preset second probability threshold, then the step of displaying the pop-up window to be recommended is executed; or, the first predicted probability and the second predicted probability are weighted and summed to obtain a fused predicted probability; if the fused predicted probability is greater than a preset third probability threshold, then the step of displaying the pop-up window to be recommended is executed.

[0085] During the process of displaying a pop-up window for recommendation based on a first predicted probability and a second predicted probability, the electronic device can determine whether the first predicted probability is greater than a preset first probability threshold, and thus determine whether the user account will close the target application within a preset time period. The electronic device can also determine whether the second predicted probability is greater than a preset second probability threshold, and thus determine whether the user account will click on the pop-up window for recommendation. When both the first predicted probability and the second predicted probability are greater than the preset first probability threshold, the electronic device determines that the user account will close the target application within the preset time period and will click on the pop-up window for recommendation. At this time, the electronic device triggers a push notification within the target application, i.e., displays the pop-up window for recommendation, thereby guiding the user account to continue accessing the content provided by the target application.

[0086] Of course, the electronic device can also perform a weighted summation of the first and second prediction probabilities to obtain a fused prediction probability. When the electronic device determines that the fused prediction probability is greater than a preset third probability threshold, the electronic device will trigger a push notification in the target application, that is, execute the display of the recommended pop-up, thereby guiding the user account to continue to access the content provided by the target application through the recommended pop-up.

[0087] Specifically, when the content corresponding to the pop-up to be recommended is a target virtual resource (e.g., a virtual red envelope pop-up), the electronic device can also respond to a claim operation on the target virtual resource by transferring the virtual resource to the virtual resource account corresponding to the user account. Specifically, the electronic device can display a claim button for the target virtual resource in the pop-up to be recommended. The user account can trigger this claim button to input a claim operation on the electronic device. After receiving the claim operation, the electronic device can instruct the server to transfer the target virtual resource to the virtual resource account corresponding to the user account. This achieves the goal of guiding the user account to continue accessing the content provided by the target application by transferring the target virtual resource to the virtual resource account corresponding to the user account.

[0088] For example, an electronic device is known to have a first predicted probability of 0.9, a second predicted probability of 0.4, a weight of 0.7 for the first predicted probability, a weight of 0.3 for the second predicted probability, and a preset third probability threshold of 0.6. The electronic device then performs a weighted sum of the first and second predicted probabilities to obtain a fused predicted probability of 0.75. Since the fused predicted probability of 0.75 is greater than the preset third probability threshold of 0.6, the electronic device triggers a push notification within the target application, displaying a pop-up window to be recommended. This pop-up window then guides the user's account to continue accessing content provided by the target application.

[0089] The technical solution of this embodiment determines whether to execute the step of displaying a pop-up window to be recommended by using multiple probability processing methods, by judging whether the first predicted probability is greater than a preset first probability threshold and whether the second predicted probability is greater than a preset second probability threshold. This enables timely use of the pop-up window to guide the user account to continue accessing the content provided by the target application, thereby improving the effective usage rate of the target application.

[0090] In an exemplary embodiment, before inputting content interaction features into the target probability prediction model to obtain user behavior prediction results, the method further includes: sending a model configuration acquisition request to a server; the model configuration acquisition request is used to instruct the server to return the model configuration parameters corresponding to the target probability prediction model; receiving the model configuration parameters returned by the server; and constructing the target probability prediction model based on the model configuration parameters.

[0091] In practical implementation, before the electronic device inputs content interaction features into the target probability prediction model to obtain user behavior prediction results, it can initialize the target probability prediction model (initialize the on-device intelligent engine). The electronic device can send a model configuration retrieval request to the server, so that the server returns the model configuration parameters corresponding to the target probability prediction model. Then, the electronic device receives the model configuration parameters and constructs the target probability prediction model based on them. For example, the client engine initialization service requests the server API at a specific time (such as a cold start of the app). The server API issues engine configuration to initialize the on-device intelligent engine. This configuration includes, for example, whether to start the on-device engine, the algorithm parameters for the on-device engine runtime, and user profiles.

[0092] The technical solution of this embodiment can achieve dynamic configuration through server-side API. Through the server-side configuration distribution, the engine parameters of the client can also be dynamically adjusted, so as to trigger the display of the recommended pop-up window in real time and accurately on the client of the electronic device.

[0093] Figure 2This is a flowchart illustrating a training method for a target probability prediction model according to an exemplary embodiment, such as... Figure 2 As shown, this training method is used in a server and includes the following steps.

[0094] In step S210, training sample data is obtained.

[0095] The training sample data includes the first training log and the second training log;

[0096] The first training log records content interaction information of the sample user account in the target application, as well as event information of the sample user account closing the target application.

[0097] The second training log records content interaction information of sample user accounts in the target application, as well as event information of sample user accounts interacting with the content corresponding to the pop-up to be recommended.

[0098] In step S220, the initial probability prediction model is trained using the training sample data to obtain the target probability prediction model.

[0099] In specific implementation, the server can use the first training log to train the first probability prediction model in the initial probability prediction model. That is, the server can input the content interaction features within a preset time period before the time to be predicted into the first probability prediction model, obtain the first output result for the time to be predicted, and query the first training log to see if there is an event where a sample user account closes the target application at the time to be predicted. Based on the query result and the first output result, the server optimizes the first probability prediction model until the optimized first probability prediction model meets the preset model training termination condition.

[0100] Specifically, the server can extract features from the content interaction characteristics within a preset time period before the time to be predicted in the first training log, obtaining content interaction sample features. Then, the server inputs these content interaction sample features into the first probability prediction model in the initial probability prediction model to obtain a first prediction output. The first prediction output includes the predicted probability that the sample user account will close the target application at the time to be predicted. Next, the server determines from the first training log whether there is a corresponding record of the sample user account closing the target application at the time to be predicted, obtaining a first target output. Based on the difference between the first target output and the first prediction output, the server uses gradient descent and backpropagation algorithms to train the first probability prediction model in the initial probability prediction model until the difference is less than a preset difference threshold. In this way, the first probability prediction model in the target probability prediction model can accurately predict the probability that the user account will close the target application at the target time based on the user account's content interaction characteristics within the target application.

[0101] In specific implementation, the server can use the second training log to train the second probability prediction model in the initial probability prediction model. That is, the server can input the content interaction features within a preset time period before the time to be predicted into the second probability prediction model, obtain the second output result for the time to be predicted, and query the second training log to see if there is an event of interaction between the sample user account and the content corresponding to the pop-up to be recommended at the time to be predicted. Based on the query result and the second output result, the second probability prediction model is optimized until the optimized second probability prediction model meets the preset model training termination condition.

[0102] Specifically, the server can extract features from content interaction information within a preset time period before the time to be predicted in the second training log, obtaining content interaction sample features. The server then inputs these content interaction sample features into the second probability prediction model within the initial probability prediction model, obtaining a second prediction output. This second prediction output includes the association probability between the sample user account and the content corresponding to the pop-up to be recommended. The server determines from the second training log whether there is a recorded event information indicating interaction between the sample user account and the content corresponding to the pop-up to be recommended at the time to be predicted, obtaining a second target output. Based on the difference between the second target output and the second prediction output, the server uses gradient descent and backpropagation algorithms to train the second probability prediction model within the initial probability prediction model until the difference is less than a preset difference threshold. In this way, the second probability prediction model within the target probability prediction model can accurately predict the association probability between the user account and the content corresponding to the pop-up to be recommended based on the user account's content interaction features within the target application.

[0103] The server can use the first probability prediction model and the second probability prediction model trained above as the target probability prediction model.

[0104] The technical solution of this embodiment uses historical content interaction features corresponding to when a sample user account clicks on a pop-up window to be recommended and historical content interaction features corresponding to when a sample user account closes the target application as training sample data to train a behavior prediction model. This allows the obtained target probability prediction model to accurately determine the first and second prediction results based on the content interaction features. This facilitates the subsequent display of the pop-up window to be recommended based on the first and second behavior prediction results. The pop-up window is then used to guide the user account to continue accessing the content provided by the target application, thereby increasing the time the user account spends accessing the target application and improving the effective usage rate of the target application.

[0105] In one exemplary embodiment, obtaining training sample data includes: obtaining content interaction logs of sample user accounts; the content interaction logs record content interaction information of sample user accounts in a target application; obtaining application closure logs of sample user accounts; the application closure logs record event information of sample user accounts closing the target application within a preset time period; and concatenating the content interaction logs and the application closure logs to obtain a first training log.

[0106] Obtain the pop-up interaction logs of the sample user accounts; the pop-up interaction logs record event information of the sample user accounts interacting with the content corresponding to the pop-up to be recommended within a preset time period; combine the content interaction logs and the pop-up interaction logs to obtain the second training logs.

[0107] For the convenience of those skilled in the art, Figure 3 This document provides a flowchart outlining the steps for obtaining training sample data. Specifically, it involves tracking user heartbeat events (reporting whether a user is still using the app every N seconds) to obtain user churn (application exit) time logs (i.e., application closure logs). It also includes tracking user pop-up click events to obtain pop-up click logs (i.e., pop-up interaction logs). The client periodically tracks on-device feature data to obtain user behavior feature logs. Based on the user churn time logs and pop-up click logs, a sample join service is used to concatenate these logs to generate a user churn training log. Similarly, concatenating the pop-up click logs with the user behavior feature logs generates a pop-up click training log. Finally, these user churn training logs and pop-up click training logs can be used as training sample data.

[0108] In practical applications, the server can also instruct electronic devices to perform the above steps to obtain training sample data and receive the training sample data returned by the electronic devices.

[0109] The technical solution of this embodiment records the content interaction characteristics of sample user accounts in the target application, the event information of sample user accounts closing the target application within a preset time period, and the event information of sample user accounts interacting with the content corresponding to the pop-up to be recommended within a preset time period, respectively, to obtain content interaction logs, application closure logs, and pop-up interaction logs. By splicing the content interaction logs and application closure logs, a first training log is obtained, and by splicing the content interaction logs and pop-up interaction logs, a second training log is obtained. This is used to train the initial probability prediction model, thereby enabling accurate subsequent mining of the correlation characteristics between the content interaction characteristics of sample user accounts and the application closure behavior and / or pop-up interaction behavior of the sample user accounts. This allows the target probability prediction model trained based on the above training data to accurately predict the application closure behavior and / or pop-up interaction behavior of the user account based on the content interaction characteristics of the user account, and then decide whether to trigger the display of the pop-up to be recommended to guide the user account to continue accessing the content provided by the target application.

[0110] Figure 4 This is a flowchart illustrating another pop-up window display method according to an exemplary embodiment, such as... Figure 3 As shown, this method is used in electronic devices and includes the following steps.

[0111] In step S410, a model configuration retrieval request is sent to the server; the model configuration retrieval request is used to instruct the server to return the model configuration parameters corresponding to the target probability prediction model.

[0112] In step S420, the model configuration parameters returned by the server are received.

[0113] In step S430, the target probability prediction model is constructed according to the model configuration parameters.

[0114] In step S440, after detecting that the target application has been launched, the content interaction features of the user account in the target application are obtained; the content interaction features are features obtained based on the interaction information between the user account and the content in the target application.

[0115] In step S450, the content interaction features are input into the target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents the probability that the user account closes the target application at a target time; the second prediction probability represents the probability of association between the user account and the content corresponding to the pop-up to be recommended.

[0116] In step S460, if the first predicted probability is greater than a preset first probability threshold and the second predicted probability is greater than a preset second probability threshold, then the pop-up window to be recommended is displayed. It should be noted that the specific limitations of the above steps can be found in the above description of the specific limitations of a pop-up window display method, and will not be repeated here.

[0117] For the convenience of those skilled in the art, Figure 5 A flowchart illustrating an application scenario for a pop-up window display method is provided. For example... Figure 5 As shown, the client (i.e., the electronic device with the target application) can initialize the target probability prediction model by calling the API interface provided by the server (i.e., the server-side). The client can send a model configuration retrieval request to the server, which will then return the model configuration parameters corresponding to the target probability prediction model. The electronic device then receives these model configuration parameters and constructs the target probability prediction model based on them, thereby initializing the on-device intelligent inference engine. Then, upon detecting the launch of the target application, the system acquires the user account's content interaction features within the target application. These content interaction features are obtained based on the interaction information between the user account and the content in the target application. The content interaction features are input into the target probability prediction model to obtain a first prediction probability and a second prediction probability. The first prediction probability represents the probability that the user account will close the target application at a target time; the second prediction probability represents the association probability between the user account and the content corresponding to the pop-up to be recommended. Based on the first and second prediction probabilities, the pop-up to be recommended is displayed. Additionally, the client can obtain feature data through the on-device event factory's feature collection service.

[0118] It should be understood that, although Figure 1 , Figure 2 and Figure 4 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 , Figure 2 and Figure 4 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0119] It is understood that the same / similar parts between the various embodiments of the methods described above in this specification can be referred to each other. Each embodiment focuses on the differences from other embodiments, and relevant parts can be referred to the description of other method embodiments.

[0120] Figure 6 This is a block diagram illustrating a pop-up window display device according to an exemplary embodiment. (Refer to...) Figure 6 The device includes:

[0121] The acquisition unit 610 is configured to acquire, upon detecting that a target application has been launched, content interaction features of a user account in the target application; the content interaction features are features obtained based on interaction information between the user account and content in the target application.

[0122] The determining unit 620 is configured to input the content interaction features into the target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents the probability that the user account closes the target application at a target time; the second prediction probability represents the probability of association between the user account and the content corresponding to the pop-up to be recommended.

[0123] Display unit 630 is configured to display the pop-up window to be recommended based on the first predicted probability and the second predicted probability.

[0124] In an exemplary embodiment, the display unit 630 is specifically configured to execute the step of displaying the pop-up window to be recommended if the first predicted probability is greater than a preset first probability threshold and the second predicted probability is greater than a preset second probability threshold; or, to perform a weighted summation of the first predicted probability and the second predicted probability to obtain a fused predicted probability; and to execute the step of displaying the pop-up window to be recommended if the fused predicted probability is greater than a preset third probability threshold.

[0125] In one exemplary embodiment, the apparatus further includes: a sending module configured to send a model configuration acquisition request to a server; the model configuration acquisition request is used to instruct the server to return model configuration parameters corresponding to the target probability prediction model; a receiving module configured to receive the model configuration parameters returned by the server; and a construction module configured to construct the target probability prediction model based on the model configuration parameters.

[0126] In one exemplary embodiment, the apparatus further includes a transfer module configured to perform a transfer of the target virtual resource to a virtual resource account corresponding to the user account in response to a claim operation on the target virtual resource.

[0127] Figure 7This is a block diagram illustrating a training apparatus for a target probability prediction model according to an exemplary embodiment. (Refer to...) Figure 7 The device includes:

[0128] The sampling unit 710 is configured to acquire training sample data; the training sample data includes a first training log and a second training log; the first training log records content interaction information of the sample user account in the target application, as well as event information of the sample user account closing the target application; the second training log records content interaction information of the sample user account in the target application, as well as event information of the sample user account interacting with the content corresponding to the pop-up to be recommended.

[0129] Training unit 720 is configured to train the initial probability prediction model using the training sample data to obtain the target probability prediction model.

[0130] In an exemplary embodiment, the training unit 720 is specifically configured to perform the following actions in the first training log: determining content interaction sample features based on content interaction features within a preset time period prior to the time to be predicted; inputting the content interaction sample features into a first probability prediction model in the initial probability prediction model to obtain a first prediction output result; the first prediction output result includes the predicted probability that the sample user account closes the target application at the time to be predicted; determining in the first training log whether there is an event information recorded at the time to be predicted that the sample user account closes the target application to obtain a first target output result; and training the first probability prediction model in the initial probability prediction model based on the difference between the first target output result and the first prediction output result until the difference is less than a preset difference threshold.

[0131] In an exemplary embodiment, the training unit 720 is specifically configured to execute in the second training log, determining content interaction sample features based on content interaction features within a preset time period before the time to be predicted; inputting the content interaction sample features into the second probability prediction model in the initial probability prediction model to obtain a second prediction output result; the second prediction output result includes the association probability between the sample user account and the content corresponding to the pop-up to be recommended; in the second training log, determining whether there is event information recording the interaction between the sample user account and the content corresponding to the pop-up to be recommended at the time to be predicted, obtaining a second target output result, and training the second probability prediction model in the initial probability prediction model based on the difference between the second target output result and the second prediction output result until the difference is less than a preset difference threshold.

[0132] In an exemplary embodiment, the sampling unit 710 is specifically configured to perform the following operations: obtaining the content interaction log of the sample user account; the content interaction log records the content interaction features of the sample user account in the target application; obtaining the application closure log of the sample user account; the application closure log records event information of the sample user account closing the target application within the preset time period; and concatenating the content interaction log and the application closure log to obtain the first training log.

[0133] In an exemplary embodiment, the sampling unit 710 is specifically configured to perform the following: obtain the pop-up interaction log of the sample user account; the pop-up interaction log records event information of the sample user account interacting with the content corresponding to the pop-up to be recommended within the preset time period; and concatenate the content interaction log and the pop-up interaction log to obtain the second training log.

[0134] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0135] Figure 8 This is a block diagram illustrating an electronic device 800 for performing the pop-up display method described above, according to an exemplary embodiment. For example, the electronic device 800 may be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, etc.

[0136] Reference Figure 8 The electronic device 800 may include one or more of the following components: processing component 802, memory 804, power supply component 806, multimedia component 808, audio component 810, input / output (I / O) interface 812, sensor component 814, and communication component 816.

[0137] Processing component 802 typically controls the overall operation of electronic device 800, such as operations associated with display, telephone calls, data communication, camera operation, and recording operations. Processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the methods described above. Furthermore, processing component 802 may include one or more modules to facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802.

[0138] Memory 804 is configured to store various types of data to support the operation of electronic device 800. Examples of such data include instructions for any application or method operating on electronic device 800, contact data, phonebook data, messages, pictures, videos, etc. Memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, optical disk, or graphene storage.

[0139] Power supply component 806 provides power to various components of electronic device 800. Power supply component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 800.

[0140] Multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 808 includes a front-facing camera and / or a rear-facing camera. When the electronic device 800 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0141] Audio component 810 is configured to output and / or input audio signals. For example, audio component 810 includes a microphone (MIC) configured to receive external audio signals when electronic device 800 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 804 or transmitted via communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

[0142] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0143] Sensor assembly 814 includes one or more sensors for providing state assessments of various aspects of electronic device 800. For example, sensor assembly 814 can detect the on / off state of electronic device 800, the relative positioning of components such as the display and keypad of electronic device 800, changes in position of electronic device 800 or its components, the presence or absence of user contact with electronic device 800, orientation or acceleration / deceleration of device 800, and temperature changes of electronic device 800. Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 814 may also include an accelerometer, gyroscope, magnetometer, pressure sensor, or temperature sensor.

[0144] Communication component 816 is configured to facilitate wired or wireless communication between electronic device 800 and other devices. Electronic device 800 can access wireless networks based on communication standards, such as WiFi, carrier networks (such as 2G, 3G, 4G, or 5G), or combinations thereof. In one exemplary embodiment, communication component 816 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 816 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0145] In an exemplary embodiment, the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0146] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 804 including instructions, which can be executed by a processor 820 of an electronic device 800 to perform the above-described method. For example, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0147] In an exemplary embodiment, a computer program product is also provided, the computer program product including instructions that can be executed by a processor 820 of an electronic device 800 to perform the above-described method.

[0148] It should be noted that the above-mentioned apparatus, electronic equipment, computer-readable storage medium, computer program product, etc., may also include other implementation methods according to the description of the method embodiments. For specific implementation methods, please refer to the description of the relevant method embodiments, which will not be elaborated here.

[0149] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0150] It should be understood that this disclosure is not limited to the precise structures 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 disclosure is limited only by the appended claims.

Claims

1. A training method for a target probability prediction model, characterized in that, The method includes: Acquire training sample data; the training sample data includes a first training log and a second training log; the first training log records the content interaction information of the sample user account in the target application, as well as the event information of the sample user account closing the target application; the second training log records the content interaction information of the sample user account in the target application, as well as the event information of the sample user account interacting with the content corresponding to the pop-up to be recommended; Using the training sample data, the initial probability prediction model is trained to obtain the target probability prediction model, including: in the first training log, determining the content interaction sample features based on the content interaction information before the time to be predicted; The content interaction sample features are input into the first probability prediction model in the initial probability prediction model to obtain a first prediction output result; the first prediction output result includes the predicted probability that the sample user account closes the target application at the time to be predicted; In the first training log, it is determined whether there is an event information of the sample user account closing the target application recorded at the time to be predicted, so as to obtain the first target output result. Based on the difference between the first target output result and the first prediction output result, the first probability prediction model in the initial probability prediction model is trained until the difference is less than a preset difference threshold. The target probability prediction model is used to determine the first prediction result and the second prediction result based on the content interaction features, so as to display the pop-up window to be recommended.

2. The training method according to claim 1, characterized in that, The step of training the initial probability prediction model using the training sample data to obtain the target probability prediction model includes: In the second training log, content interaction sample features are determined based on content interaction information prior to the time to be predicted; The content interaction sample features are input into the second probability prediction model in the initial probability prediction model to obtain the second prediction output result; the second prediction output result includes the association probability between the sample user account and the content corresponding to the pop-up to be recommended; In the second training log, it is determined whether there is an event information recording the interaction between the sample user account and the content corresponding to the pop-up to be recommended at the time to be predicted, so as to obtain the second target output result. Based on the difference between the second target output result and the second prediction output result, the second probability prediction model in the initial probability prediction model is trained until the difference is less than the preset difference threshold.

3. The training method according to claim 1, characterized in that, The acquisition of training sample data includes: Obtain the content interaction log of the sample user account; the content interaction log records the content interaction information of the sample user account in the target application. Obtain the application shutdown log of the sample user account; the application shutdown log records event information of the sample user account closing the target application within a preset time period; The first training log is obtained by concatenating the content interaction log and the application closing log.

4. The training method according to claim 1, characterized in that, The method further includes: Obtain the pop-up interaction log of the sample user account; the pop-up interaction log records event information of the sample user account interacting with the content corresponding to the pop-up to be recommended within a preset time period; The second training log is obtained by concatenating the content interaction log and the pop-up interaction log.

5. A pop-up window display method, characterized in that, include: After detecting that the target application has been launched, the content interaction characteristics of the user account in the target application are obtained; The content interaction feature is a feature obtained based on the interaction information between the user account and the content in the target application; The content interaction features are input into the target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents the probability that the user account closes the target application at a target time; the second prediction probability represents the association probability between the user account and the content corresponding to the pop-up to be recommended; the target probability prediction model is trained according to the training method of the target probability prediction model as described in any one of claims 1 to 4; The pop-up window to be recommended is displayed based on the first predicted probability and the second predicted probability.

6. The pop-up window display method according to claim 5, characterized in that, The step of displaying the pop-up window to be recommended based on the first predicted probability and the second predicted probability includes: If the first predicted probability is greater than a preset first probability threshold and the second predicted probability is greater than a preset second probability threshold, then the step of displaying the pop-up window to be recommended is executed. or, The first predicted probability and the second predicted probability are weighted and summed to obtain the fused predicted probability; if the fused predicted probability is greater than a preset third probability threshold, then the step of displaying the pop-up window to be recommended is executed.

7. The pop-up window display method according to claim 5, characterized in that, The method further includes: A model configuration retrieval request is sent to the server; the model configuration retrieval request is used to instruct the server to return the model configuration parameters corresponding to the target probability prediction model; Receive the model configuration parameters returned by the server; The target probability prediction model is constructed based on the model configuration parameters.

8. The pop-up window display method according to claim 5, characterized in that, If the content corresponding to the pop-up to be recommended is a target virtual resource, after the step of displaying the pop-up to be recommended based on the first prediction probability and the second prediction probability, the method further includes: In response to the claim operation of the target virtual resource, the target virtual resource is transferred to the virtual resource account corresponding to the user account.

9. A training device for a target probability prediction model, characterized in that, include: The sampling unit is configured to acquire training sample data. The training sample data includes a first training log and a second training log; The first training log records content interaction information of the sample user account in the target application, as well as event information of the sample user account closing the target application; The second training log records content interaction information of sample user accounts in the target application, as well as event information of the interaction between the sample user accounts and the content corresponding to the pop-up to be recommended; The training unit is configured to train the initial probability prediction model using the training sample data to obtain the target probability prediction model; The training unit is specifically configured to execute in the first training log to determine content interaction sample features based on content interaction information before the time to be predicted; input the content interaction sample features into the first probability prediction model in the initial probability prediction model to obtain a first prediction output result; the first prediction output result includes the predicted probability that the sample user account closes the target application at the time to be predicted; in the first training log, determine whether there is an event information recorded at the time to be predicted that the sample user account closes the target application to obtain a first target output result, and train the first probability prediction model in the initial probability prediction model based on the difference between the first target output result and the first prediction output result until the difference is less than a preset difference threshold; The target probability prediction model is used to determine the first prediction result and the second prediction result based on the content interaction features, so as to display the pop-up window to be recommended.

10. The training device according to claim 9, characterized in that, The training unit is specifically configured to execute in the second training log to determine content interaction sample features based on content interaction features before the time to be predicted; input the content interaction sample features into the second probability prediction model in the initial probability prediction model to obtain a second prediction output result; the second prediction output result includes the association probability between the sample user account and the content corresponding to the pop-up to be recommended; in the second training log, determine whether there is event information recording the interaction between the sample user account and the content corresponding to the pop-up to be recommended at the time to be predicted, obtain a second target output result, and train the second probability prediction model in the initial probability prediction model based on the difference between the second target output result and the second prediction output result until the difference is less than a preset difference threshold.

11. The training device according to claim 9, characterized in that, The sampling unit is specifically configured to retrieve the content interaction log of the sample user account; the content interaction log records the content interaction information of the sample user account in the target application. Obtain the application closure log of the sample user account; the application closure log records event information of the sample user account closing the target application within a preset time period; concatenate the content interaction log and the application closure log to obtain the first training log.

12. The training device according to claim 9, characterized in that, The sampling unit is specifically configured to perform the following: obtain the pop-up interaction log of the sample user account; the pop-up interaction log records event information of the sample user account interacting with the content corresponding to the pop-up to be recommended within a preset time period; and concatenate the content interaction log and the pop-up interaction log to obtain the second training log.

13. A pop-up display device, characterized in that, include: The acquisition unit is configured to acquire content interaction features of a user account in the target application after detecting that the target application has been launched; The content interaction feature is a feature obtained based on the interaction information between the user account and the content in the target application; The determining unit is configured to input the content interaction features into a target probability prediction model to obtain a first prediction probability and a second prediction probability; the first prediction probability represents the probability that the user account closes the target application at a target time; the second prediction probability represents the association probability between the user account and the content corresponding to the pop-up to be recommended; the target probability prediction model is trained according to the training method of the target probability prediction model as described in any one of claims 1 to 4. The display unit is configured to display the pop-up window to be recommended based on the first predicted probability and the second predicted probability.

14. The pop-up display device according to claim 13, characterized in that, The display unit is specifically configured to execute the step of displaying the pop-up window to be recommended if the first predicted probability is greater than a preset first probability threshold and the second predicted probability is greater than a preset second probability threshold; or, to perform a weighted summation of the first predicted probability and the second predicted probability to obtain a fused predicted probability; and to execute the step of displaying the pop-up window to be recommended if the fused predicted probability is greater than a preset third probability threshold.

15. The pop-up display device according to claim 13, characterized in that, The device further includes: The sending module is configured to send a model configuration retrieval request to the server; the model configuration retrieval request is used to instruct the server to return the model configuration parameters corresponding to the target probability prediction model; the receiving module is configured to receive the model configuration parameters returned by the server; the building module is configured to build the target probability prediction model according to the model configuration parameters.

16. The pop-up display device according to claim 13, characterized in that, If the content corresponding to the pop-up window to be recommended is a target virtual resource, the device further includes: a transfer module configured to perform a request operation on the target virtual resource and transfer the target virtual resource to the virtual resource account corresponding to the user account.

17. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the training method as described in any one of claims 1 to 4, and / or the pop-up display method as described in any one of claims 5 to 8.

18. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the training method as described in any one of claims 1 to 4, and / or the pop-up display method as described in any one of claims 5 to 8.

19. A computer program product, the computer program product comprising instructions, characterized in that, When the instruction is executed by the processor of the electronic device, the electronic device is able to perform the training method as described in any one of claims 1 to 4, and / or the pop-up display method as described in any one of claims 5 to 8.

Images