Information push method, computer device and computer program product

Abstract

The present application discloses an information push method, a computer device and a computer program product, the information push method comprising: obtaining the click probability of candidate push contents pushed by target users for different candidate applications; based on the click probability, determining a target application matching the target user from the candidate applications, and determining a target push content matching the target user from the candidate push contents; based on the manufacturer push channel, pushing the target push content to the target application in the electronic device corresponding to the target user, so that the electronic device displays the target push content through the target application. The method can improve the push effect of information push.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and more specifically, to an information push method, a computer device, and a computer program product. Background Art

[0002] With the rapid advancement of technology and living standards, electronic devices (such as smart phones, tablet computers, etc.) have become one of the commonly used electronic products in people's lives. In addition, with the development of Internet technology, products are often pushed to users through applications installed in electronic devices. However, the current method is not accurate enough in pushing information, resulting in poor push effects. Summary of the invention

[0003] The present application proposes an information push method, a computer device and a computer program product, which can improve the push effect of information push.

[0004] In a first aspect, an embodiment of the present application provides an information push method, characterized in that the method includes: obtaining the click probability of candidate push content pushed by the target user for different candidate applications; based on the click probability, determining a target application matching the target user from the candidate applications, and determining a target push content matching the target user from the candidate push content; based on the manufacturer push channel, pushing the target push content to the target application in the electronic device corresponding to the target user, so that the electronic device displays the target push content through the target application.

[0005] In the second aspect, an embodiment of the present application provides an information push device, which includes: a probability prediction module, a content determination module and a content push module, wherein the probability prediction module is used to obtain the click probability of the candidate push content pushed by the target user for different candidate applications; the content determination module is used to determine the target application matching the target user from the candidate applications based on the click probability, and determine the target push content matching the target user from the candidate push content; the content push module is used to push the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer push channel, so that the electronic device displays the target push content through the target application.

[0006] In a third aspect, an embodiment of the present application provides a computer device comprising: one or more processors; a memory; one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to execute the information push method provided in the first aspect above.

[0007] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a program code is stored. The program code can be called by a processor to execute the information push method provided in the first aspect above.

[0008] In a fifth aspect, an embodiment of the present application provides a computer program product, which includes a computer program, and the computer program is stored in a storage medium; a processor of a computer device reads the computer program from the storage medium, and the processor executes the computer program, so that the computer device executes the information push method provided in the first aspect above.

[0009] The solution provided by the present application obtains the click probability of the candidate push content pushed by the target user for different candidate applications, determines the target application matching the target user from the candidate applications based on the click probability, and determines the target push content matching the target user from the candidate push content, and pushes the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer's push channel, so that the electronic device displays the target push content through the target application. As a result, when pushing information, the push effect of the user for different applications and push content (that is, the click probability of the user for the candidate push content pushed by different applications) can be combined to determine the content to be pushed and the application to display the push content, so that the accuracy of information push can be improved, thereby improving the user's reach rate of the push content, and then improving the push effect of information push. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0011] Figure 1 A schematic diagram of an application environment provided by an embodiment of the present application is shown.

[0012] Figure 2 A schematic flow chart of an information push method according to an embodiment of the present application is shown.

[0013] Figure 3 A flow chart of an information push method according to another embodiment of the present application is shown.

[0014] Figure 4 A flow chart of an information push method according to another embodiment of the present application is shown.

[0015] Figure 5A schematic diagram showing the principle of generating a candidate user set provided by an embodiment of the present application is shown.

[0016] Figure 6 A flow chart of an information push method according to another embodiment of the present application is shown.

[0017] Figure 7 A schematic diagram showing the principle of the information push method provided in an embodiment of the present application is shown.

[0018] Figure 8 A schematic diagram showing the principle of the information push method provided in an embodiment of the present application is shown.

[0019] Fig. 9 A block diagram of an information push device according to an embodiment of the present application is shown.

[0020] Fig.10 It is a block diagram of a computer device according to an embodiment of the present application for executing the information push method according to an embodiment of the present application.

[0021] Fig.11 It is a storage unit of an embodiment of the present application for storing or carrying program codes for implementing the information push method according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0023] With the development of Internet technology and the development and application of big data technology, Internet advertising has become one of the important means of acquiring customers for advertisers. From the perspective of advertisers, advertising is usually carried out in two ways: passive request and active reach. In the passive request method, users need to open the corresponding application (Application, APP) and perform corresponding operations in the APP before the electronic device can request the advertiser's advertisement. The user who opens the APP used to push advertisements may not be the advertiser's full target user or part of the target user, resulting in a low user reach rate and poor advertising push effect; in the active reach method, Internet users receive a large number of social, marketing, and information push messages every day, resulting in message fatigue. Users often delete or do not read messages in batches, which makes the effective exposure rate of advertisements low when pushing advertisements through active reach. In addition, the premise that the user terminal can receive the APP's in-site push message is that the APP is running in the background. If the user's frequency of using the APP is low, it will also affect the reach and exposure rate of the pushed content.

[0024] In response to the above problems, the inventors have proposed the information push method, computer device, and computer program product provided in the embodiments of the present application, which can combine the user's push effects for different applications and push content (i.e., the user's click probability for candidate push content pushed by different applications) to determine the content to be pushed and the application that displays the push content when pushing information, thereby improving the accuracy of information push, thereby improving the user's reach rate for push content, and further improving the push effect of information push. Among them, the specific information push method is described in detail in the subsequent embodiments.

[0025] The following first introduces the scenarios involved in the embodiments of the present application.

[0026] like Figure 1 As shown, in Figure 1 The scene shown includes a server 100 and an electronic device 200. The server 100 may be a server (network access server), a server cluster (cloud server) consisting of several servers, or a cloud computing center (database server). The electronic device 200 may be any electronic device with communication and storage functions, including but not limited to a PC (Personal Computer), a PDA (tablet computer), a smart TV, a smart phone, a smart wearable device, or other smart communication devices with network connection functions.

[0027] The server 100 and the electronic device 200 can interact through the network to receive or send information. In the embodiment of the present application, the server 100 can determine the click probability of the target user corresponding to the electronic device 200 for the candidate push content pushed by different candidate applications, and then determine the target application matching the target user from the candidate applications based on the click probability of the candidate push content pushed by different candidate applications, and determine the target push content matching the target user from the candidate push content, and finally push the target push content to the target application in the electronic device 200 based on the manufacturer push channel, so that the target application in the electronic device 200 can display the received target push content, and push the above target push content through the manufacturer push method.

[0028] The information push method provided in the embodiment of the present application is described in detail below with reference to the accompanying drawings.

[0029] See also Figure 2 , Figure 2 The following is a flow chart of an information push method provided by an embodiment of the present application. Figure 2 The process shown is described in detail, and the information push method may specifically include the following steps:

[0030] Step S110: Obtain the click probability of the target user for the candidate push contents pushed by different candidate applications.

[0031] In an embodiment of the present application, when a computer device pushes information to a target user, it can obtain the click probability of the target user for the candidate push content pushed by different candidate applications. The above candidate applications can be applications that can push information through manufacturer push (Push); the above candidate push content can include various types of push content, for example, it can include content used to promote and publicize products, services or brands, and for example, it can include social push content, and for example, it can include information push content, etc. The candidate push content can be used to attract the attention of the target audience, convey specific marketing information, and prompt the audience to take action (such as purchase, registration, learn more information, etc.). The above click probability represents the probability of a user clicking to view the candidate push content.

[0032] Among them, manufacturer push refers to the push of a long link maintained by the manufacturer of the electronic device, which is a system-level channel. After the APP accesses the manufacturer push service, even when the APP process is killed (automatic cleaning by the system, active cleaning by the user, etc.), the corresponding push information can be sent to the APP in the user's electronic device through the manufacturer push channel. Most of the information is notification bar information, and then the APP application can be opened by clicking on the notification bar information to process the information. In the embodiment of the present application, after receiving the push content, the candidate application can display the push content, thereby achieving the effect of the candidate application pushing the received push content.

[0033] It is understandable that for candidate push content, when the same content is pushed through different applications, the probability of a user clicking on the pushed content is different. Therefore, the user's click probability on the candidate push content can be predicted based on the situation where different candidate push content is pushed by different candidate applications, so as to more accurately determine the application and push content for information push to the target user.

[0034] In some embodiments, the computer device may predict the click probability of the target user for the candidate pushed content pushed by different candidate applications based on the pre-trained prediction model for predicting click probability and the relevant information of the target user.

[0035] Among them, the above prediction models can be click-through rate prediction models (CTRM), conversion rate prediction models (CVRM), user behavior sequence models (UBSM), etc. The specific type of prediction model may not be limited in the embodiments of the present application. The click-through rate prediction model is a very important type of model in the fields of recommendation systems and online advertising, which is used to predict the probability of users clicking on pushed content (such as advertisements, search results, recommended products, etc.); the conversion rate prediction model is a model used to predict the probability of users clicking on specific behaviors (such as purchases, registrations, downloads, etc.); the user behavior sequence model is a model used to analyze and predict the behavior sequence of users in a specific environment, and can output the probability of the behavior occurring.

[0036] Exemplarily, the candidate push contents include push content 1, push content 2, and push content 3, and the candidate applications include application 1, application 2, and application 3. The click probability of each candidate push content when pushed by different candidate applications can be shown in the following table:

[0037]

[0038]

[0039] In the above table, probability a represents the target user's click probability for push content 1 pushed through application 1, probability b represents the target user's click probability for push content 2 pushed through application 1, probability b represents the target user's click probability for push content 3 pushed through application 1, probability d represents the target user's click probability for push content 1 pushed through application 2, probability g represents the target user's click probability for push content 1 pushed through application 2, and so on.

[0040] In a possible implementation, considering that push effects of different candidate applications are different, different prediction models may be used to predict click probabilities for different candidate applications. When obtaining the click probabilities of candidate push contents pushed by target users for different candidate applications, the prediction model corresponding to each candidate application may be used, and based on the relevant information of the target user, the click probability of the target user for the candidate push contents pushed by each candidate application may be predicted.

[0041] Optionally, for the case where each candidate push content is pushed by a different candidate application, the above relevant information of the target user can be input into the prediction model corresponding to each candidate application to obtain the result output by the prediction model corresponding to each candidate application. The result output by the prediction model corresponding to each candidate application is the click probability of the candidate push content being pushed by the target user for each candidate application.

[0042] Of course, in the embodiments of the present application, the specific method of obtaining the above click probability may not be limited. For example, the above click probability may be determined by separately obtaining the probability of users using the candidate applications and the probability of users clicking to view the candidate pushed content, and then further determining the above click probability based on the probability of users using the candidate applications and the probability of users clicking to view the candidate pushed content.

[0043] Step S120: Based on the click probability, determine a target application that matches the target user from the candidate applications, and determine a target push content that matches the target user from the candidate push content.

[0044] In an embodiment of the present application, after determining the click probability of the target user for the candidate push content pushed by different candidate applications, the application used to display the push content and the push content that needs to be pushed when the information is pushed to the target user this time can be determined. Based on the above click probability, the computer device can determine the target application that matches the target user from the candidate applications, and determine the target push content that matches the target user from the candidate push content. It can be understood that the above click probability can reflect the push effect to a certain extent, and the click probability corresponds to the candidate application and the candidate push content, so the above target application and target push content can be determined based on the above click probability.

[0045] In some embodiments, when a computer device determines candidate push content and candidate applications that match a target user based on the click probabilities of candidate push content pushed by the target user for different candidate applications, it can determine the click probability that satisfies the target probability condition from the above click probabilities; then, the candidate application corresponding to the click probability that satisfies the target probability condition is determined as the above target application, and the candidate push content corresponding to the click probability that satisfies the target probability condition is determined as the above target push content.

[0046] In a possible implementation, the above target probability condition may include that the click probability is greater than a probability threshold. That is, the computer device may determine a click probability greater than a probability threshold from the click probabilities of candidate push contents pushed by the target user for different candidate applications, and determine the candidate application corresponding to the click probability greater than the probability threshold as the above target application, and determine the candidate push content corresponding to the click probability greater than the probability threshold as the above target push content.

[0047] In a possible implementation, the above target probability condition may include that the click probability is the maximum value. That is, the computer device may determine the maximum click probability from the click probabilities of the candidate push contents pushed by the target user for different candidate applications, and determine the candidate application corresponding to the maximum click probability as the above target application, and determine the candidate push content corresponding to the maximum click probability as the above target push content.

[0048] Step S130: based on the manufacturer push channel, pushing the target push content to the target application in the electronic device corresponding to the target user, so that the electronic device displays the target push content through the target application.

[0049] In the embodiment of the present application, after determining the target push content and target application that match the target user, the computer device can push the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer push channel. Correspondingly, the target application in the electronic device corresponding to the target user can receive the target push content sent by the computer device, so that the target push content can be displayed through the target application, realizing the push method through the manufacturer, and pushing the above target push content through the target application.

[0050] Among them, the above-mentioned manufacturer push channel can be a long connection channel established between the computer device and the electronic device of the target user. The computer device can push messages to the APP in the electronic device in real time through the long connection channel. Regardless of whether the APP process in the electronic device exists, the message can be received normally.

[0051] In some implementations, after receiving the above target push content, the target application in the target user's electronic device can display a notification message in the notification bar based on the target push content, and the notification message displayed in the notification bar can include information about the target push content.

[0052] The information push method provided in the embodiment of the present application can, when pushing information, combine the push effects of users for different applications and push contents (i.e., the click probability of users for candidate push contents pushed by different applications) to determine the content to be pushed and the application that displays the push contents, thereby improving the accuracy of information push, thereby improving the user's reach rate of the push contents, and further improving the push effect of information push; secondly, in the embodiment of the present application, information push is performed by manufacturer push, so that even if the application is not running in the background, the pushed content can reach the user's electronic device, thereby improving the reach rate of the pushed content; and, when information push is performed by manufacturer push, when applied to advertising push, advertisers can connect to advertising media management (i.e., the management of candidate applications for push information) once, and can achieve push by different applications, which can reduce advertisers' time cost, development cost, labor cost and other costs; in addition, when content push is performed by manufacturer push, different applications can be selected for push for push content, thereby achieving multi-application collaboration and further improving the reach rate of push content.

[0053] See also Figure 3 , Figure 3 A schematic diagram of a flow chart of an information push method provided by another embodiment of the present application is shown. Figure 3 The process shown is described in detail, and the information push method may specifically include the following steps:

[0054] Step S210: predicting a first probability that the target user uses different candidate applications, and predicting a second probability that the target user clicks on candidate push content.

[0055] In an embodiment of the present application, when the computer device obtains the click probability of the target user for the candidate push content pushed by different candidate applications, it can predict the probability of the target user using different candidate applications (as the first probability), and predict the probability of the target user clicking on the candidate push content (as the second probability), so as to determine the click probability of the target user for the candidate push content pushed by the candidate application based on the obtained first probability and second probability. It can be understood that the above first probability can reflect the user's preference or liking for the candidate application, and the second probability can reflect the user's interest in the candidate recommended content, so the above first probability and the second probability can be obtained, so as to determine the above click probability based on the above first probability and the second probability.

[0056] In some embodiments, the computer device may obtain historical usage data of the target user using the application, for example, the application usage log may be obtained from the electronic device corresponding to the target user; based on the historical usage data of the application, the first probability of the user using each candidate application may be determined. The above historical usage data may include the number of historical usages of the application, the historical usage duration, the average usage duration, the usage frequency, etc. It can be understood that the historical usage data of the application may reflect the user's usage preferences and habits for the application, so the probability of the user using the candidate application may be determined based on the above historical usage data.

[0057] In a possible implementation, when determining the first probability of a user using each candidate application based on the historical usage data of the application, the above historical usage data can be input into a pre-trained usage probability prediction model, and the probability value output by the usage probability prediction model can be obtained as the above first probability. Among them, the usage probability prediction model can be a deep learning model (Deep Learning Models), a logistic regression model (Logistic Regression Models), etc., and the specific type of the usage probability prediction model may not be limited in the embodiments of the present application.

[0058] In one possible implementation, the above historical usage data may include the target user's historical usage counts and historical usage duration for each candidate application. When determining the above first probability for each candidate application, the historical usage counts and historical usage duration may be weighted calculated, such as weighted summing, based on the first weight corresponding to the historical usage counts and the second weight corresponding to the historical usage duration; then, the value obtained after the weighted calculation is normalized to obtain the above first probability. Exemplarily, the historical usage count is T1, the historical usage duration is T2, the first weight is a, the second weight is b, and the sum of a and b is 1. A weighted summation formula may be used to perform the weighted summation, and the weighted summation formula is: a*T1+b*T2. Then, the value obtained by the weighted summation is normalized by the normalization formula to obtain the above first probability, where the normalization formula is

[0059] x * =xx min

[0060] Can be: x max -x min , where x * represents the first probability corresponding to the currently calculated candidate application (i.e., the probability that the user uses the candidate application), x is the weighted sum of the above values ​​obtained for the currently calculated candidate application, and x max is the maximum value determined from the above weighted sum values ​​obtained from all applications used by the user, xmin It is the minimum value determined from the above weighted sum values ​​obtained from all applications used by the user.

[0061] In some embodiments, considering that users in the same region may have similar preferences and habits when using applications, the historical usage data of all users in the region where the target user is located can be obtained. Based on the historical usage data of all users, the average number of times and average usage duration of each application by users in the region can be determined; based on the average number of times and average usage duration of each application, the above first probability corresponding to each candidate application can be determined.

[0062] In a possible implementation, when determining the above first probability for each candidate application, the average number of uses and the average usage duration may be weighted according to the third weight corresponding to the above average number of uses and the fourth weight corresponding to the average usage duration, such as weighted summation, etc. Then, the value obtained after the weighted calculation is normalized to obtain the above first probability. For example, the average number of uses is T3, the average usage duration is T4, the third weight is c, the second weight is d, and the sum of c and d is 1. The weighted summation formula can be used to perform the weighted summation, and the weighted summation formula is: c*T3+d*T4. Then, the value obtained by the weighted summation is normalized by the normalization formula to obtain the above first probability, where the normalization formula can be: Among them, x * represents the first probability corresponding to the currently calculated candidate application (i.e., the probability that the user uses the candidate application), x is the weighted sum of the above values ​​obtained for the currently calculated candidate application, and x max is the maximum value determined from the above weighted sum values ​​obtained for all applications, x min is the minimum value determined from the above weighted summed values ​​obtained for all applications.

[0063] In some embodiments, when predicting the second probability that the target user clicks on the candidate pushed content, the computer device can use the first prediction model and based on the user characteristics of the target user, the content characteristics of the candidate pushed content, and the device characteristics of the electronic device to predict the second probability that the target user clicks on the candidate pushed content.

[0064] In a possible implementation, user characteristics may include, but are not limited to, basic information, historical behavior, and real-time interaction data of the user. Basic information may include age, gender, occupation, hobbies, etc. Historical behavior may include browsing behavior (e.g., browsing history, dwell time, click behavior, etc. of the user on a website or application), search behavior (search keywords, search frequency, search time, etc. of the user), purchase behavior (purchase history, purchase preference, purchase frequency, purchase amount, etc. of the user), and usage habits (time, frequency, manner, etc. of the user using a product or service). Real-time interaction data may include click data, browsing data, search data, form data, evaluation data, consultation data, geographic location, information of the currently used device, an identifier of the user session, etc.

[0065] Among them, the above basic information can be determined from the registration information found by the computer device according to the target user's account; for the above real-time interactive data, click data can be a record of users clicking on elements such as pages, links, buttons, etc., including click time, click location, click frequency, etc.; browsing data can be the user's browsing trajectory on the page, including browsing time, scrolling depth, page dwell time, etc.; search data can be the keywords entered by the user in the search box, search time, search result selection, etc.; form data can be the information entered by the user when filling in the form, such as registration information, feedback, etc.; evaluation data can be the user's evaluation, rating, comments, etc. on the product or service; consultation data can be consultations and questions raised by users through online customer service, message boards, social media and other channels; geographic location can be the geographic location information when the user visits; information on the currently used device can be the type of device used by the user, operating system, browser version, etc.; the identifier of the user session can be a unique identifier used to identify the user session, which can track the user's behavior throughout the session.

[0066] In a possible implementation, the content features of the candidate push content may include features such as the type, text, and material of the push content. The computer device may pre-store content features for each candidate push content, and when the second probability needs to be determined, the computer device may read the content features corresponding to the current candidate push content.

[0067] In a possible implementation, the device features of the above electronic devices may include device contextual information, which may refer to background information related to the user's current behavior, needs or status, such as time context information, space context information, social context information, device context information, task context information, etc. In this implementation, the contextual information may include the current time, device information, network status, etc.

[0068] In a possible implementation, when training the above first prediction model, since the task of the first prediction model is to predict the probability of clicking on the pushed content, a deep neural network (DNN) can be selected. Of course, the first prediction model can also be other binary classification models; when obtaining sample data, the positive sample is the pushed content that is pushed and clicked by the user, and the negative sample is the pushed content that is pushed but not clicked by the user. For the positive and negative samples, the above user features, content features and device features can be obtained, and the user features, content features and device features are used as sample features, and the sample features are labeled with the labels of the positive samples or the labels of the negative samples, so that the sample data for training the above deep neural network can be obtained; according to the sample data, the above deep neural network is trained to obtain the above first prediction model. When measuring the performance of the trained first prediction model, the model classification performance can be measured by the receiver operating characteristic (ROC) curve and the area under the curve (Area Under the Curve, AUC) value.

[0069] In the above implementation, when the above deep neural network is trained according to the sample data, the sample features can be input into the deep neural network to obtain the output result of the deep neural network; then the loss value can be determined according to the difference between the output result and the label marked by the sample features; after the loss value is determined, the deep neural network can be iteratively trained according to the loss value to obtain the final first prediction model. When the deep neural network is iteratively trained according to the loss value, the model parameters of the deep neural network can be adjusted according to the calculated loss value; the sample features are repeatedly input into the initial model to obtain the output result of the deep neural network, and the loss value can be determined according to the output result and the label marked by the sample features, as well as the model parameters of the deep neural network according to the calculated loss value, until the training end condition is met to obtain the trained first prediction model. The training end condition of iterative training may include: the number of iterative training reaches the target number; or the total loss value of the deep neural network meets the set condition.

[0070] It should be noted that the training of the first prediction model can be carried out in advance, and then each time the second probability needs to be determined, the trained first prediction model can be used for prediction, without the need to train the first prediction model each time the second probability needs to be determined.

[0071] Step S220: Based on the first probability and the second probability, determine the click probability of the target user for the candidate push content pushed by different candidate applications, and the click probability is positively correlated with the first probability and the second probability.

[0072] In an embodiment of the present application, after determining the first probability that the target user uses different candidate applications, and the second probability that the target user clicks on the candidate push content, the click probability of the candidate push content pushed by the target user for different candidate applications can be determined based on the above first probability and the second probability. It can be understood that the first probability corresponds to a candidate application, and the second probability corresponds to a candidate push content, so based on the first probability and the second probability, the probability related to the candidate application and the candidate push content (i.e., the above click probability) can be determined. Moreover, the click probability is positively correlated with the first probability and the second probability, that is, when the second probability remains unchanged, the larger the first probability, the larger the determined click probability; when the first probability remains unchanged, the larger the second probability, the larger the determined click probability.

[0073] In some embodiments, when determining the click probability of the target user for the candidate push content pushed by the candidate application based on the first probability and the second probability, the product of the first probability and the second probability can be obtained, and the obtained product is used as the above click probability. Among them, after obtaining the click probability in the above manner by combining the first probabilities corresponding to different candidate applications and the second probabilities corresponding to different candidate push contents, the click probability of the target user for different candidate push contents pushed by the candidate application can be obtained.

[0074] In some embodiments, when determining the click probability of the target user for the candidate pushed content pushed by the candidate application based on the first probability and the second probability, the first probability and the second probability can also be weighted calculated based on the weight corresponding to the first probability and the weight corresponding to the second probability, such as weighted sum, and the value obtained after the weighted calculation is used as the above click probability.

[0075] Of course, in the embodiment of the present application, the specific implementation method of determining the above click probability based on the above first probability and the second probability may not be limited.

[0076] In some implementations, after obtaining the above click probabilities, the click probabilities obtained this time can be calibrated based on the click probabilities obtained for the user history to improve the accuracy of the obtained click probabilities, thereby further improving the accuracy of the subsequently determined push content and applications.

[0077] Step S230: Based on the click probability, determine a target application that matches the target user from the candidate applications, and determine a target push content that matches the target user from the candidate push content.

[0078] Step S240: Pushing the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer push channel, so that the electronic device displays the target push content through the target application.

[0079] In the embodiment of the present application, step S230 and step S240 can refer to the contents of other embodiments and will not be repeated here.

[0080] The information push method provided in the embodiment of the present application can, when pushing information, combine the user's push effects for different applications and push content (i.e., the user's click probability for candidate push content pushed by different applications) to determine the content that needs to be pushed and the application that displays the push content, thereby improving the accuracy of information push, thereby improving the user's reach rate for push content, and further improving the push effect of information push; in addition, the embodiment of the present application also provides a method for determining the click probability of candidate push content pushed by different applications of users, by predicting a first probability that the target user uses different candidate applications, and predicting a second probability that the target user clicks on the candidate push content, and then determining the above click probability based on the first probability and the second probability, thereby ensuring that the determined click probability is related to the user's preference for using applications and the user's interest in the push content, thereby further improving the accuracy of the determined content that needs to be pushed and the application that displays the push content.

[0081] See also Figure 4 , Figure 4 A schematic diagram of a flow chart of an information push method provided by another embodiment of the present application is shown. Figure 4 The process shown is described in detail, and the information push method may specifically include the following steps:

[0082] Step S310: Determine the third probability that each user clicks on the pushed content in different time periods through the second prediction model and based on the user characteristics corresponding to each user, the time characteristics of different time periods, the application characteristics of the candidate application and the device characteristics of each user.

[0083] In an embodiment of the present application, the third probability of each user clicking on the pushed content in different time periods can be determined for all users who need to push content, so as to push information to each user in a suitable time period and improve the user's reach rate of the pushed content. Among them, the computer device can use the second prediction model, and based on the user characteristics corresponding to each user, the time characteristics of different time periods, the application characteristics of the candidate application, and the device characteristics of each user, determine the third probability of each user clicking on the pushed content in different time periods. Among them, the above different time periods can be time periods divided for 1 day, for example, each hour of the 24 hours in 1 day can be divided into 1 time period.

[0084] In some embodiments, time features may be time-related features, which may include hours, days of the week, months, and user activity in each time period; user features may include but are not limited to basic information about users, click rates of pushed content by users in the last N days, etc.; application features may include arrival data, exposure data, click data, etc. of content pushed historically by candidate applications; device features may include device usage, such as the number of launches per day, usage duration, etc.

[0085] In some embodiments, when training the above second prediction model, since the task of the second prediction model is to predict the probability of a user clicking on the pushed content in different time periods, a multi-classification model can be selected as the second prediction model, for example, a Deep Factorization Machine (DeepFM) model. DeepFM can automatically learn feature interactions when dealing with multi-classification problems, avoiding complex artificial feature engineering. The model integrates deep learning and factorization machines, taking into account both nonlinear and linear features, and can efficiently process sparse data to improve prediction accuracy. When obtaining positive samples, the time when each user clicks on the pushed content can be mapped to the corresponding time period, and the top N time periods with the most clicks on the pushed content by each user (i.e., the number of clicks in the N time periods is greater than the number of clicks in other time periods) are obtained, and by random sampling, the sample magnitude of each time period is controlled to be consistent with the proportion of the total positive samples; when obtaining positive samples, the time when each user clicks on the pushed content can be mapped to the corresponding time period, and the N time periods with the least clicks on the pushed content by each user (i.e., the number of clicks in the N time periods is less than the number of clicks in other time periods) are obtained, and by random sampling, the sample magnitude of each time period is controlled to be consistent with the proportion of the total negative samples. For positive samples and negative samples, the above user features, time features, application features and device features can be obtained, and the user features, time features, application features and device features are used as sample features, and the sample features are labeled with positive sample labels or negative sample labels, so that sample data for training the model can be obtained; according to the sample data, the above multi-classification model is trained to obtain the above second prediction model. When measuring the performance of the trained second prediction model, the performance of the model can be measured according to the Group Area Under the Curve (GAUC) indicator. The advantage of GAUC is that it can distinguish the performance differences between groups and the whole, ensuring that for each user, the predicted click-through rates in each time period are in order.

[0086] In the above implementation, when the above multi-classification model is trained according to sample data, the sample features can be input into the multi-classification model to obtain the output result of the multi-classification model; then the loss value can be determined according to the difference between the output result and the label marked by the sample features; after the loss value is determined, the multi-classification model can be iteratively trained according to the loss value to obtain the final second prediction model. When the multi-classification model is iteratively trained according to the loss value, the model parameters of the multi-classification model can be adjusted according to the calculated loss value; the sample features are repeatedly input into the initial model to obtain the output result of the multi-classification model, and the loss value can be determined according to the output result and the label marked by the sample features, as well as the model parameters of the multi-classification model according to the calculated loss value, until the training end condition is met to obtain the trained second prediction model. The training end condition of iterative training may include: the number of iterative training reaches the target number; or the total loss value of the multi-classification model meets the set condition.

[0087] It should be noted that the training of the second prediction model can be carried out in advance, and then the trained second prediction model can be used for prediction each time the third probability needs to be determined, without the need to train the second prediction model each time the third probability needs to be determined.

[0088] Step S320: Based on the third probability, determine candidate user sets corresponding to different time periods.

[0089] In the embodiment of the present application, after determining the third probability of each user clicking on the pushed content in different time periods, the candidate user sets corresponding to the different time periods can be determined based on the third probability of each user clicking on the pushed content in different time periods. In other words, each user is divided into the corresponding time period.

[0090] In some embodiments, considering that the number of push content items that a user of a candidate application receives every day is limited, the N time periods corresponding to each user can be determined based on the above third probability and the limited number of push content items, where N matches the limited number and the third probability corresponding to the N time periods is greater than the third probability corresponding to other time periods; based on the N time periods corresponding to each user, a set of candidate users corresponding to different time periods is generated.

[0091] In the above implementation, for each user, the third probability of the user in each time period can be sorted in order from large to small, and then the first N time periods are determined from the sorting results, and the N time periods are the time periods corresponding to the user; according to the N time periods corresponding to each user, the users in the same time period are divided into the user set corresponding to the time period, so that the candidate user set corresponding to different time periods can be obtained. The candidate user set obtained above can also be understood as a request set, that is, a set of requests for push content by users in the candidate user set. For example, the third probability predicted for user A is ranked from high to low as the third probability corresponding to the time period at 8 o'clock, the third probability corresponding to the time period at 10 o'clock, the third probability corresponding to the time period at 14 o'clock, the third probability corresponding to the time period at 16 o'clock..., and the above N is 3, then user A can be divided into the candidate user set corresponding to the time period at 8 o'clock, the candidate user set corresponding to the time period at 10 o'clock, and the candidate user set corresponding to the time period at 14 o'clock.

[0092] It can be understood that through the above implementation, each user can be in the user set with the third higher probability time period above him, so that when information is pushed subsequently, the user's reach rate of pushed content can be further improved by pushing information in the time period corresponding to the user set to which he belongs.

[0093] For example, see Figure 5 The computer device can predict the third probability based on the user characteristics corresponding to each user, the time characteristics of different time periods, the application characteristics of the candidate applications and the device characteristics of each user, and then divide each user into sets corresponding to different time periods according to the limit number N, so as to obtain candidate user sets corresponding to different time periods.

[0094] Step S330: determining a candidate user set corresponding to a target time period from the candidate user sets corresponding to the different time periods, and determining a target user from the candidate user set corresponding to the target time period.

[0095] In the embodiment of the present application, the above target time period may be any time period, and the target user may be any user in the candidate user set corresponding to the target time period.

[0096] Step S340: Obtain the click probability of the target user for the candidate push contents pushed by different candidate applications.

[0097] Step S350: Based on the click probability, determine a target application that matches the target user from the candidate applications, and determine a target push content that matches the target user from the candidate push content.

[0098] In the embodiment of the present application, steps S340 to S350 may refer to the contents of other embodiments and will not be described in detail here.

[0099] Step S360: In the target time period, based on the manufacturer push channel, the target push content is pushed to the target application in the electronic device corresponding to the target user, so that the electronic device displays the target push content through the target application.

[0100] In the embodiment of the present application, when information is pushed to the above target user, after the above target application and the target push content are determined, the target push content can be pushed to the target application in the electronic device corresponding to the target user in the above target time period based on the manufacturer push channel. Since the above target time period is a time period determined according to the above third probability corresponding to the target user, the probability of the target user clicking on the push content in the target time period is relatively high, so the above method can further improve the user's reach rate of the push content.

[0101] In some embodiments, when multiple target applications and target push content are determined for the above target users, the target push content can be pushed to the above target applications based on the manufacturer push channel multiple times within the target time period, thereby achieving the push of the target push content through each of the above multiple target applications respectively.

[0102] In a possible implementation, in order to prevent the user's electronic device from receiving a large amount of push content at the same time point within the target time period, the target push content can be pushed to the above target application based on the manufacturer's push channel every target interval time within the target time period.

[0103] In some implementations, the determined push content can be pushed to the electronic device of each user in the candidate user set corresponding to each time period according to the sequence of all time periods through the above implementations, thereby realizing information push for all users.

[0104] In a possible implementation, before information is pushed to each user based on the candidate user set corresponding to each time period, the corresponding users can be filtered out from each candidate user set based on a pre-set screening rule. The screening rule can be a targeting or restriction rule corresponding to the candidate push content. For example, when the candidate push content includes financial credit advertisements, the screening rule can include only users in second-tier and above cities, restricted age groups, credit scores, and not pushing to users with bad loan records. By filtering users in the candidate user set according to the screening rules above, crowd targeting can be achieved, further reducing the number of candidate users, and also reducing the cost of information push.

[0105] The information push method provided in the embodiment of the present application can, when pushing information, combine the push effects of users for different applications and pushed content (i.e., the click probability of users for candidate pushed content pushed by different applications) to determine the content that needs to be pushed and the application that displays the pushed content, thereby improving the accuracy of information push, thereby improving the user's reach rate for pushed content, and further improving the push effect of information push; in addition, the time period for information push to users is determined based on the predicted probability of users clicking on pushed content in different time periods, thereby further improving the user's reach rate for pushed content, and also avoiding message push to all users in the same time period, thereby reducing the processing pressure and cost of computer equipment.

[0106] See also Figure 6 , Figure 6 A schematic diagram of a flow chart of an information push method provided by another embodiment of the present application is shown. Figure 6 The process shown is described in detail, and the information push method may specifically include the following steps:

[0107] Step S410: Obtain the click probability of the target user for the candidate push contents pushed by different candidate applications.

[0108] In the embodiment of the present application, step S410 can refer to the contents of other embodiments and will not be repeated here.

[0109] Step S420: Based on the target threshold of the click probability, determine the target probability from the click probability, and determine the candidate application corresponding to the target probability as the target application, and determine the candidate push content corresponding to the target probability as the target push content, and the target threshold is determined by proportional integral differential PID control.

[0110] In an embodiment of the present application, when determining a target application matching a target user from candidate applications based on the above click probability, and determining a target push content matching a target user from candidate push content, considering that when information is pushed in a manufacturer push manner, the manufacturer usually charges in a cost per click (CPC) manner. In order to take into account the manufacturer's revenue from information push and not push too much content to users, generally speaking, the actual daily click-through rate is expected to reach a certain value. Therefore, the expected click-through rate (as a target threshold of the click probability) can be determined by proportional-integral-derivative (PID) control, and based on the determined target threshold of the click probability, the target probability is determined from the click probabilities of the candidate push contents pushed for different candidate applications, and the candidate application corresponding to the target probability is determined as the target application, and the candidate push content corresponding to the target probability is determined as the target push content.

[0111] Among them, the PID control method is used to adjust the output of the system to the desired target value. PID control is achieved through three parameters: proportion (P), integral (I) and differential (D). In the embodiment of the present application, PID control can dynamically adjust the sending level of the pushed content to ensure that the actual click-through rate reaches the expected target.

[0112] In some implementations, the computer device may dynamically adjust the target threshold of the click probability based on PID control every day. The target threshold CTR_t+1 to be determined on day T+1 is calculated using the following formula:

[0113]

[0114] Wherein, error_t = CTR_t target – CTR_t true, CTR_t true can be obtained based on the click rate monitored on the tth day, and CTR_t target is the target threshold determined on the tth day.

[0115] In the above implementation, the computer device can monitor the push content sent to the user's electronic device every day to monitor the push data, click data, etc. of the push content. Of course, the monitored data can also include other data. For example, in the scenario of advertising push, the advertiser's budget consumption ratio can also be monitored.

[0116] In some embodiments, when determining the above target probability, the click probability with a click probability greater than the above target threshold can be determined from the click probability of the candidate push content pushed by the target user for different candidate applications, and the determined click probability is used as the above target probability. It can be understood that when the target threshold is different, the number of target probabilities determined above is also different, so the number of combinations of target applications and target push content finally determined is also different, and the number of times the subsequent computer device pushes content to the target user's electronic device is also different, and thus the sending level of the pushed content can be dynamically adjusted, so that when the computer device pushes information through the manufacturer push method, it can take into account the manufacturer's information push income and not push too much content to the user.

[0117] Step S430: based on the manufacturer push channel, push the target push content to the target application in the electronic device corresponding to the target user, so that the electronic device displays the target push content through the target application.

[0118] In the embodiment of the present application, step S430 can refer to the contents of other embodiments and will not be repeated here.

[0119] In some embodiments, each candidate push content can be managed and distributed in a computer device, and when pushing the candidate push content to an application in an electronic device, the requirements of different applications for the format and size of the candidate push content can be adapted so that the corresponding application in the electronic device can display the push content normally after receiving the push content.

[0120] In some implementations, the target pushed content pushed to the target application in the electronic device may carry a dwell time to instruct the target application of the electronic device to cancel display of the target pushed content after the display time reaches the dwell time.

[0121] The information push method provided in the embodiment of the present application can, when pushing information, combine the push effects of users for different applications and push content (i.e., the click probability of users for candidate push content pushed by different applications) to determine the content that needs to be pushed and the application that displays the push content, thereby improving the accuracy of information push, thereby improving the user's reach rate for push content, and further improving the push effect of information push; in addition, according to the threshold of the click probability determined by PID control, and according to the click probability of users for candidate push content pushed by different applications, the content that needs to be pushed and the application that displays the push content are determined, so that when the computer device pushes information through the manufacturer push method, it can take into account the manufacturer's income from information push and not push too much content to the user.

[0122] Next, we will combine the information push scenario of advertising push with Figure 7 and Figure 8 The information push method involved in the above embodiment is illustrated by examples.

[0123] See also Figure 7 and Figure 8 In the information push scenario of advertising push, the time period in which the probability of users clicking on the pushed content is relatively high can be determined through the time period division module, and based on the determined time period, candidate user sets corresponding to different time periods are generated; when information is pushed to users in each candidate user set, the users in the candidate user set are filtered based on the screening rules through the multi-advertisement multi-application optimization sorting module, and then the first probability of users using different candidate applications is predicted, and the second probability of users clicking on the candidate pushed content is predicted, and the click probability of the user for the candidate pushed content pushed by different candidate applications is determined based on the first probability and the second probability; the target threshold of the click probability is determined through the traffic intelligent adjustment module and based on the PID control method, and then the target probability that meets the conditions is determined from the click probability determined according to the target threshold, and the application corresponding to the target probability is determined as the target application, and the push content corresponding to the target probability is determined as the target push content; then, the target application and the target push content are determined through the multi-application message display module, and the content is pushed by the manufacturer; in addition, the application set composed of applications for content push can also be managed through the application advertising media management module, and the number of restrictions N mentioned in the aforementioned embodiment, the type of content pushed by the application in the application set, etc. can also be managed.

[0124] See also Fig. 9 , which shows a structural block diagram of an information push device 500 provided in an embodiment of the present application. The information push device 500 applies the above-mentioned electronic device, and the information push device 500 includes: a probability prediction module, a content determination module and a content push module, wherein the probability prediction module is used to obtain the click probability of the candidate push content pushed by the target user for different candidate applications; the content determination module is used to determine the target application matching the target user from the candidate applications based on the click probability, and determine the target push content matching the target user from the candidate push content; the content push module is used to push the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer push channel, so that the electronic device displays the target push content through the target application.

[0125] In some embodiments, the probability prediction module can be specifically used to: predict a first probability that the target user uses different candidate applications, and predict a second probability that the target user clicks on the candidate push content; based on the first probability and the second probability, determine the click probability, and the click probability is positively correlated with the first probability and the second probability.

[0126] In a possible implementation, the probability prediction module may be specifically configured to obtain the product of the first probability and the second probability as the click probability.

[0127] In a possible implementation, the probability prediction module may also be used to predict a second probability that the target user clicks on the candidate pushed content through a first prediction model and based on user characteristics of the target user, content characteristics of the candidate pushed content, and device characteristics of the electronic device.

[0128] In some embodiments, the content determination module can be specifically used to: determine the target probability from the click probability based on the target threshold of the click probability, and determine the candidate application corresponding to the target probability as the target application, and determine the candidate push content corresponding to the target probability as the target push content, and the target threshold is determined by proportional integral differential PID control.

[0129] In some embodiments, the target user is determined from a set of candidate users corresponding to the target time period, and the information push device may further include a set generation module and a user determination module. The probability prediction module may also be used to determine the third probability of each user clicking on the pushed content in different time periods through a second prediction model and based on the user characteristics corresponding to each user, the time characteristics of different time periods, the application characteristics of the candidate application, and the device characteristics of each user; the set generation module is used to determine the candidate user sets corresponding to different time periods based on the third probability; the user determination module is used to determine the candidate user set corresponding to the target time period from the candidate user sets corresponding to the different time periods, and determine the target user from the candidate user set corresponding to the target time period.

[0130] In a possible implementation, the content push module may be specifically configured to: push the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer push channel during the target time period.

[0131] In a possible implementation, the set generation module can be specifically used to: determine N time periods corresponding to each user based on the third probability and the limited number of pushed content, where N matches the limited number, and the third probability corresponding to the N time periods is greater than the third probability corresponding to other time periods; and generate candidate user sets corresponding to different time periods based on the N time periods corresponding to each user.

[0132] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0133] In several embodiments provided in the present application, the coupling between modules may be electrical, mechanical or other forms of coupling.

[0134] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or software functional modules.

[0135] In summary, the solution provided by the present application obtains the click probability of the candidate push content pushed by the target user for different candidate applications, determines the target application matching the target user from the candidate applications based on the click probability, and determines the target push content matching the target user from the candidate push content, and pushes the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer's push channel, so that the electronic device displays the target push content through the target application. As a result, when pushing information, it is possible to combine the push effect of the user for different applications and push content (that is, the click probability of the user for the candidate push content pushed by different applications) to determine the content to be pushed and the application to display the push content, thereby improving the accuracy of information push, thereby improving the user's reach rate of push content, and then improving the push effect of information push.

[0136] Please refer to Fig.10 , which shows a structural block diagram of a computer device provided in an embodiment of the present application. The computer device 100 may be a server. The computer device 100 in the present application may include one or more of the following components: a processor 110, a memory 120, and one or more application programs, wherein the one or more application programs may be stored in the memory 120 and configured to be executed by one or more processors 110, and the one or more application programs are configured to execute the method described in the aforementioned method embodiment.

[0137] The processor 110 may include one or more processing cores. The processor 110 uses various interfaces and lines to connect various parts of the entire computer device 100, and executes various functions and processes data of the computer device 100 by running or executing instructions, programs, code sets or instruction sets stored in the memory 120, and calling data stored in the memory 120. Optionally, the processor 110 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 110 can integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 110, but may be implemented separately through a communication chip.

[0138] The memory 120 may include a random access memory (RAM) or a read-only memory (ROM). The memory 120 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 120 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the following various method embodiments, etc. The data storage area may also store data (such as a phone book, audio and video data, chat record data) created by the computer device 100 during use.

[0139] Please refer to Fig.11 , which shows a structural block diagram of a computer-readable storage medium provided in an embodiment of the present application. The computer-readable medium 800 stores program codes, which can be called by a processor to execute the method described in the above method embodiment.

[0140] The computer readable storage medium 800 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk, or a ROM. Optionally, the computer readable storage medium 800 includes a non-transitory computer-readable storage medium. The computer readable storage medium 800 has storage space for program code 810 that performs any method step in the above method. These program codes can be read from or written to one or more computer program products. The program code 810 can be compressed, for example, in an appropriate form.

[0141] An embodiment of the present application also provides a computer program product, which includes a computer program, and the computer program is stored in a storage medium; a processor of a computer device reads the computer program from the storage medium, and the processor executes the computer program, so that the computer device executes the information push method provided in the above embodiment.

[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An information push method, characterized in that: The method comprises: Obtain the click probability of the target user for the candidate push content pushed by different candidate applications; Based on the click probability, determining a target application matching the target user from the candidate applications, and determining a target push content matching the target user from the candidate push content; Based on the manufacturer push channel, the target push content is pushed to the target application in the electronic device corresponding to the target user, so that the electronic device displays the target push content through the target application.

2. The method according to claim 1, characterized in that The step of obtaining the click probability of the target user for the candidate pushed content pushed by different candidate applications includes: Predicting a first probability that the target user uses different candidate applications, and predicting a second probability that the target user clicks on the candidate pushed content; The click probability is determined based on the first probability and the second probability, and the click probability is positively correlated with the first probability and the second probability.

3. The method according to claim 2, characterized in that The determining the click probability based on the first probability and the second probability includes: The product of the first probability and the second probability is obtained as the click probability.

4. The method according to claim 2, characterized in that: The predicting the second probability that the target user clicks the candidate pushed content includes: The second probability that the target user clicks on the candidate pushed content is predicted by the first prediction model based on the user characteristics of the target user, the content characteristics of the candidate pushed content, and the device characteristics of the electronic device.

5. The method according to claim 1, characterized in that The step of determining a target application matching the target user from the candidate applications based on the click probability, and determining a target push content matching the target user from the candidate push content includes: Based on the target threshold of the click probability, the target probability is determined from the click probability, and the candidate application corresponding to the target probability is determined as the target application, and the candidate push content corresponding to the target probability is determined as the target push content, and the target threshold is determined by proportional integral differential PID control.

6. The method according to any one of claims 1 to 5, characterized in that: The target user is determined from a set of candidate users corresponding to the target time period. Before obtaining the click probability of the target user for the candidate pushed content pushed by different candidate applications, the method further includes: Determining, by the second prediction model and based on user characteristics corresponding to each user, time characteristics of different time periods, application characteristics of the candidate application, and device characteristics of each user, a third probability that each user clicks on the pushed content in different time periods; Based on the third probability, determining candidate user sets corresponding to different time periods; A candidate user set corresponding to the target time period is determined from the candidate user sets corresponding to the different time periods, and the target user is determined from the candidate user set corresponding to the target time period.

7. The method according to claim 6, characterized in that The pushing of the target push content to the target application in the electronic device corresponding to the target user based on the manufacturer push channel includes: In the target time period, based on the manufacturer push channel, the target push content is pushed to the target application in the electronic device corresponding to the target user.

8. The method according to claim 6, characterized in that The determining, based on the third probability, candidate user sets corresponding to different time periods includes: Based on the third probability and the limited number of pushed content, determine N time periods corresponding to each user, where N matches the limited number of pushed content, and the third probability corresponding to the N time periods is greater than the third probability corresponding to other time periods; Based on the N time periods corresponding to the respective users, candidate user sets corresponding to different time periods are generated.

9. A computer device, characterized in that: include: one or more processors; Memory; One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the method according to any one of claims 1 to 8.

10. A computer program product, characterized in that The computer program product comprises a computer program, wherein the computer program is stored in a storage medium; The processor of the computer device reads the computer program from the storage medium, and the processor executes the computer program, so that the computer device executes the method according to any one of claims 1 to 8.

Images