Display information pushing method and device based on user browsing frequency and electronic equipment

Abstract

The embodiment of the invention relates to a display information pushing method and device based on user browsing frequency and electronic equipment. The method comprises the following steps: acquiring behavior data corresponding to a target user; based on the browsing behavior data, determining browsing frequency information of the target user; determining a user type to which the target user belongs based on the browsing frequency information; performing display value prediction on the behavior data by using a value prediction model corresponding to the user type to obtain a display value score of the target user; and if the display value score meets the preset display condition, sending target display information to a terminal used by the target user. According to the embodiment of the invention, the method achieves the selection of the corresponding value prediction model according to the browsing frequency of the user, achieves the prediction of the display value of the user, achieves the prediction of the display value through different value prediction models for different types of user groups, improves the precision of value prediction, and improves the user experience. And therefore, the display information can be more accurately put to the user.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, electronic device, and computer-readable storage medium for pushing display information based on user browsing frequency. Background Art

[0002] When delivering display information, the provider of display information will pre-evaluate the target users from multiple angles, such as information pulling, usage duration (for example, audio and video playback duration), and benefits obtained from users. Based on the evaluation results, it will be determined whether it is necessary to deliver display information to the target users.

[0003] However, product usage frequency varies significantly among different users. For example, some users launch applications more frequently than others. Using a unified modeling approach to assess the value of advertising across all users makes it difficult to predict value for different user groups. Therefore, to improve the targeting of information delivery, it is necessary to accurately assess the value of different user groups. Summary of the Invention

[0004] In view of this, in order to solve some or all of the above technical problems, the embodiments of the present application provide a method, device, electronic device and computer-readable storage medium for pushing display information based on user browsing frequency.

[0005] In a first aspect, an embodiment of the present application provides a method for pushing display information based on user browsing frequency, the method comprising: obtaining behavioral data corresponding to a target user, wherein the behavioral data comprises browsing behavior data of the target user for a target product, and viewing data recorded when the target user performs related operations based on display information related to the target product; determining the browsing frequency information of the target user based on the browsing behavior data; determining the user type to which the target user belongs based on the browsing frequency information; extracting a value prediction model corresponding to the user type; using the value prediction model, performing display value prediction on the behavioral data to obtain a display value score of the target user, wherein the display value score represents the probability of the target user performing a preset operation after viewing the target display information; and sending the target display information to the terminal used by the target user if the display value score meets the preset display conditions.

[0006] In one possible implementation, extracting a value prediction model corresponding to a user type includes: if the user type indicates that the target user is a short silent period user, extracting a first value prediction model corresponding to the short silent period user; using the value prediction model, performing display value prediction on the behavior data to obtain a display value score of the target user, including: obtaining a preset first identifier indicating that display information is delivered to the target user, and a second identifier indicating that display information is not delivered to the target user; merging the first identifier with the behavior data to obtain first behavior data, and merging the second identifier with the behavior data to obtain second behavior data; using the first value prediction model, performing display value prediction on the first behavior data and the second behavior data, respectively, to obtain a first display value score and a second display value score; determining the display value score of the target user based on the difference between the first display value score and the second display value score.

[0007] In one possible embodiment, the first value prediction model is pre-trained according to the following steps: obtaining first sample behavior data recorded after display information is delivered to users with a short silent period, and second sample behavior data recorded without display information, wherein the first sample behavior data has corresponding first labeling information, and the second sample behavior data has corresponding first labeling information; using the first sample behavior data and the corresponding first identifier as the input of the first initial value prediction model, and the first labeling information as the expected output of the initial value prediction model, to train the first initial value prediction model; using the second sample behavior data and the corresponding second identifier as the input of the initial value prediction model, and the second labeling information as the expected output of the initial value prediction model, to train the first initial value prediction model; in response to the first initial value prediction model meeting the training end condition, determining the first initial value prediction model as the first value prediction model.

[0008] In one possible implementation, extracting a value prediction model corresponding to the user type includes: if the user type indicates that the target user is a long silent period user, extracting a second value prediction model corresponding to the long silent period user; using the value prediction model, performing display value prediction on the behavioral data to obtain the display value score of the target user, including: using the second value prediction model, performing display value prediction on the behavioral data to obtain the display value score of the target user.

[0009] In one possible embodiment, the second value prediction model is pre-trained according to the following steps: obtaining third sample behavior data recorded for users with long silent periods, wherein the third sample behavior data has corresponding third annotation information; using the third sample behavior data as the input of the second initial value prediction model, and using the third annotation information as the expected output of the second initial value prediction model, to train the second initial value prediction model; in response to the second initial value prediction model meeting the training end conditions, determining the second initial value prediction model as the second value prediction model.

[0010] In one possible implementation, the value prediction model includes a multi-task shared network and at least two value prediction networks; using the value prediction model, the behavior data is subjected to display value prediction to obtain a display value score of the target user, including: using the multi-task shared network to extract features of the behavior data to obtain multi-task basic features; using each value prediction network in the at least two value prediction networks to perform value prediction on the multi-task basic features to obtain at least two display value prediction values; and fusing the at least two display value prediction values to obtain a display value score.

[0011] In one possible embodiment, at least two value prediction networks include a usage time prediction network and at least one revenue prediction network; using each value prediction network in the at least two value prediction networks, value prediction is performed on multi-task basic features to obtain at least two display value prediction values, including: using each revenue prediction network in the at least one revenue prediction network to perform revenue value prediction on the multi-task basic features respectively, and obtain at least one revenue prediction value as the display value prediction value corresponding to the at least one revenue prediction network, wherein the revenue prediction value represents the amount of revenue obtained from the target user using the target display information within a first preset time period in the future; using the usage time prediction network, usage time prediction is performed on the multi-task basic features to obtain the predicted usage time of the target user, wherein the predicted usage time represents the length of time the target user uses the target website within a second preset time period in the future.

[0012] In a second aspect, an embodiment of the present application provides a display information push device based on user browsing frequency, the device comprising: an acquisition module for acquiring behavior data corresponding to a target user, wherein the behavior data comprises browsing behavior data of the target user for a target product, and viewing data recorded when the target user performs related operations based on display information related to the target product; a first determination module for determining the browsing frequency information of the target user based on the browsing behavior data; a second determination module for determining the user type to which the target user belongs based on the browsing frequency information; an extraction module for extracting a value prediction model corresponding to the user type; a prediction module for using the value prediction model to perform display value prediction on the behavior data to obtain a display value score of the target user, wherein the display value score represents the probability of the target user performing a preset operation after viewing the target display information; and a sending module for sending the target display information to the terminal used by the target user if the display value score meets the preset display conditions.

[0013] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a memory for storing a computer program; a processor for executing the computer program stored in the memory, and when the computer program is executed, it implements the method of any embodiment of the method for pushing display information based on user browsing frequency of the above-mentioned first aspect of the present application.

[0014] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method of any embodiment of the method for pushing display information based on user browsing frequency as described in the first aspect above is implemented.

[0015] In a fifth aspect, an embodiment of the present application provides a computer program, which includes a computer-readable code. When the computer-readable code runs on a device, the processor in the device implements a method of any embodiment of the method for pushing display information based on user browsing frequency as described in the first aspect above.

[0016] The method, device, electronic device, and computer-readable storage medium for pushing display information based on user browsing frequency provided in the embodiments of the present application obtain behavioral data corresponding to the target user, determine the user's browsing frequency information based on the browsing behavior data included in the behavioral data, and use the value prediction model corresponding to the browsing frequency information to predict the display value of the behavioral data to obtain the display value score of the target user. If the display value score meets the preset display conditions, the target display information is sent to the terminal used by the target user. The embodiments of the present application select a corresponding value prediction model based on the user's browsing frequency to predict the user's display value, thereby realizing the use of different value prediction models for display value prediction for different types of user groups, improving the accuracy of value prediction, and thus delivering display information to users more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0019] One or more embodiments are exemplarily illustrated by pictures in the corresponding drawings. These exemplifications do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings are represented as similar elements. Unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0020] Figure 1 A flowchart of a method for pushing display information based on user browsing frequency provided in an embodiment of the present application;

[0021] Figure 2 A flowchart of another method for pushing display information based on user browsing frequency provided in an embodiment of the present application;

[0022] Figure 3 A flowchart of a training method for a first value prediction model provided in an embodiment of the present application;

[0023] Figure 4 A flowchart of another method for pushing display information based on user browsing frequency provided in an embodiment of the present application;

[0024] Figure 5 A flowchart of a training method for a second value prediction model provided in an embodiment of the present application;

[0025] Figure 6 A schematic diagram of the structure of the value prediction model provided in the embodiment of the present application;

[0026] Figure 7 A flowchart of another method for pushing display information based on user browsing frequency provided in an embodiment of the present application;

[0027] Figure 8 A schematic diagram of the structure of another value prediction model provided in an embodiment of the present application;

[0028] Figure 9 A flowchart of another method for pushing display information based on user browsing frequency provided in an embodiment of the present application;

[0029] Figure 10 A schematic diagram of the structure of a device for pushing display information based on user browsing frequency provided in an embodiment of the present application;

[0030] Figure 11 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0031] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It is apparent that the described embodiments are only a portion of the embodiments of the present application, rather than all of the embodiments. It should be noted that, unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions, and numerical values described in these embodiments do not limit the scope of the present application.

[0032] Those skilled in the art will understand that the terms "first" and "second" in the embodiments of the present application are only used to distinguish between different steps, devices, modules and other objects, and neither represent any specific technical meaning nor indicate the logical order between them.

[0033] It should also be understood that in this embodiment, “a plurality of” may refer to two or more than two, and “at least one” may refer to one, two or more than two.

[0034] It should also be understood that any component, data or structure mentioned in the embodiments of the present application can generally be understood as one or more, unless explicitly limited or otherwise indicated in the context.

[0035] In addition, the term "and / or" in this application is simply a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this application generally indicates that the related objects are in an "or" relationship.

[0036] It should also be understood that the description of each embodiment in this application focuses on the differences between the embodiments, and the same or similar aspects can be referenced with each other. For the sake of brevity, they will not be described one by one.

[0037] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present disclosure, its application, or uses.

[0038] Technologies, circuits, and devices known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, the above-mentioned technologies, circuits, and devices should be considered part of the specification.

[0039] It should be noted that like reference numerals and letters refer to like items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0040] It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of this application can be combined with each other. To facilitate understanding of the embodiments of this application, the application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0041] In order to solve the technical problem that the existing technology does not perform different display value predictions for users with different browsing frequencies, resulting in low accuracy of display information push, the present application provides a display information push method based on user browsing frequency, which can use different value prediction models to predict display value for different types of user groups, thereby improving the accuracy of value prediction and the targetedness of display information push.

[0042] Figure 1 A flow chart of a method for pushing display information based on user browsing frequency provided in an embodiment of the present application. This method can generally be applied to a server, by setting a value prediction model on the server, analyzing the user's behavior data, and then pushing display information to the user terminal. In addition, the execution subject of this method can be hardware or software. When the above-mentioned execution subject is hardware, the execution subject can be one or more of the above-mentioned electronic devices. For example, a single electronic device can execute this method, or a plurality of electronic devices can cooperate with each other to execute this method. When the above-mentioned execution subject is software, this method can be implemented as multiple software or software modules, or as a single software or software module. No specific limitation is made here.

[0043] like Figure 1As shown, the method specifically includes:

[0044] Step 101: Obtain behavior data corresponding to the target user.

[0045] In some embodiments, the behavioral data includes browsing behavior data of the target user for a target product. The target product can be a physical product or a virtual product. For example, the target product can be an application, a website, or a product sold on a shopping platform. The target user can browse the information provided by the target product. During browsing, the various behaviors performed by the target user can be recorded in real time to obtain behavioral data. For example, if the target product is a video playback application, the behavioral data may include data such as the program startup time, program shutdown time, video playback duration, and video playback type.

[0046] The aforementioned behavioral data also includes viewing data recorded when the target user performs related operations based on the display information related to the target product. The display information is used to display relevant information about the target product to the target user. The target user can perform various operations based on the display information, and the data recorded for these operations is the viewing data. For example, the display information is advertising information, the target product is a video playback application, and the target user views the display information on other applications. Based on the display information, the target user can click on the display information and jump to the video playback application, or perform operations such as purchases and membership registration. Therefore, the recorded viewing data can represent advertising exposure, advertising clicks, advertising revenue, and other situations.

[0047] Step 102: Determine the browsing frequency information of the target user based on the browsing behavior data.

[0048] In some embodiments, browsing frequency information indicates how frequently the target user browses the information provided by the target product. For example, if the target product is an application, the browsing frequency information may include the time each application is launched. Based on the time each application is launched, information such as the maximum interval duration and average interval duration of the user's use of the application can be determined. Any of these information can be used as browsing frequency information.

[0049] Step 103: Determine the user type of the target user based on the browsing frequency information.

[0050] In some embodiments, user types can be pre-classified based on how frequently each user browses information about the target product. For example, based on browsing frequency information, users can be classified into long-silent-period users (i.e., users whose time interval between two uses of the target product is greater than or equal to a preset duration) and short-silent-period users (i.e., users whose time interval between two uses of the target product is less than a preset duration).

[0051] Step 104: extract the value prediction model corresponding to the user type.

[0052] In some embodiments, corresponding value prediction models may be pre-set for different user types.

[0053] For example, we can pre-acquire behavioral data from multiple users with long silent periods as sample data, annotate the expected output data, and use this data to train a value prediction model for users with long silent periods. We can also pre-acquire behavioral data from multiple users with short silent periods as sample data, annotate the expected output data, and use this data to train a value prediction model for users with short silent periods.

[0054] The value prediction model is pre-trained using a machine learning algorithm. Specifically, a large number of training samples are collected in advance. Each training sample includes sample behavior data, as well as annotated information about the revenue earned from the sample users and the duration of their use of the target product. During model training, the sample behavior data serves as the model input, and the annotated revenue and usage duration information serves as the model's expected output. By adjusting the model parameters, the error between the model's actual and expected outputs is minimized, resulting in a trained value prediction model.

[0055] Step 105: Use the value prediction model to predict the display value of the behavior data to obtain the display value score of the target user.

[0056] In some embodiments, the display value score represents the probability that the target user will perform a preset action after viewing the target display information. That is, the higher the display value score, the greater the probability that the user will perform the preset action, and the higher the potential profit for the provider of the target display information.

[0057] The target display information may include various types of information, such as advertising information, recommendation information, item description information, etc. The preset operations include various types of pre-set operations, such as click operations, purchase operations after clicks, registration operations, etc.

[0058] Step 106: If the display value score meets the preset display conditions, the target display information is sent to the terminal used by the target user.

[0059] In some embodiments, the aforementioned display conditions are pre-set, representing criteria for determining whether a target user can potentially generate revenue for the provider of the target display information. For example, a display value score threshold may be set; if the display value score is greater than or equal to the threshold, the target display information may be pushed to the target user. For another example, the display value scores of multiple users may be tallied and ranked in descending order of display value scores. If the target user is in the top X users, the target display information may be pushed to the target user.

[0060] The method for pushing display information based on user browsing frequency provided in an embodiment of the present application obtains the behavior data corresponding to the target user, determines the browsing frequency information of the user based on the browsing behavior data included in the behavior data, and uses the value prediction model corresponding to the browsing frequency information to predict the display value of the behavior data to obtain the display value score of the target user. If the display value score meets the preset display conditions, the target display information is sent to the terminal used by the target user. The embodiment of the present application realizes the selection of a corresponding value prediction model based on the user's browsing frequency to predict the user's display value, thereby realizing the display value prediction for different types of user groups using different value prediction models, improving the accuracy of value prediction, and thus delivering display information to users more accurately.

[0061] In some optional implementations, such as Figure 2 As shown, step 104 includes:

[0062] Step 1041: If the user type indicates that the target user is a short silent period user, extract a first value prediction model corresponding to the short silent period user.

[0063] Among them, the first value prediction model is a model pre-trained for users with short silence periods.

[0064] The step 105 includes:

[0065] Step 1051: Obtain a preset first flag indicating that the display information is delivered to the target user, and a second flag indicating that the display information is not delivered to the target user.

[0066] The first identifier and the second identifier can each be a constant, or a vector composed of constants. For example, the first value prediction model can be pre-built according to the S-Learner modeling paradigm, with the first identifier represented by treatment = 1 and the second identifier represented by treatment = 0.

[0067] Step 1052: Merge the first identifier and the behavior data to obtain first behavior data, and merge the second identifier and the behavior data to obtain second behavior data.

[0068] The first identifier or the second identifier may be merged with the behavior data by directly merging the first identifier or the second identifier as a feature into the behavior data. Alternatively, the first identifier or the second identifier may be calculated (e.g., by multiplying the first identifier or the second identifier by some data in the behavior data) and then merging the calculated data with the behavior data.

[0069] Step 1053: Use the first value prediction model to predict the display value of the first behavior data and the second behavior data respectively to obtain a first display value score and a second display value score.

[0070] Specifically, the first behavior data and the second behavior data can be sequentially input into the first value prediction model. A first display value score corresponding to the first behavior data and a second display value score corresponding to the second behavior data are obtained. The first display value score represents the potential revenue obtained from the target user if the target display information is delivered to the target user. The second display value score represents the potential revenue obtained from the target user if the target display information is not delivered to the target user.

[0071] Step 1054: Determine the exhibition value score of the target user based on the difference between the first exhibition value score and the second exhibition value score.

[0072] The difference between the first display value score and the second display value score represents the display value gain, that is, the increase in the display value score when the target display information is pushed to the target user compared to when the target display information is not pushed to the target user. The higher the increase, the greater the possibility of obtaining potential benefits from the target user. This difference can be used as the display value score of the target user, or the difference can be converted (for example, normalized or converted to a percentage system) to obtain the display value score.

[0073] By setting the first and second identifiers, this embodiment can predict the gain in display value for both the case where the target display information is pushed to the target user and the case where the target display information is not pushed to the target user. Because users with short silent periods frequently use the target product and are less sensitive to the target display information, the above gain can be calculated for both the case where the display information is delivered and the case where the display information is not delivered. This allows the resulting display value score to more accurately match the target user's habits of using the target product, improving the accuracy of determining the user who will receive the display information.

[0074] In some optional implementations, such as Figure 3 As shown, the first value prediction model is pre-trained according to the following steps:

[0075] Step 301: Acquire first sample behavior data recorded after display information is delivered to a user with a short silent period, and second sample behavior data recorded without display information.

[0076] Among them, the first sample behavior data has corresponding first annotation information, and the second sample behavior data has corresponding first annotation information. The first annotation information and the second annotation information can respectively include information on revenue obtained from users with short silent periods, information on the duration of use of the target product by users with short silent periods (for example, when the target product is a video playback application, the duration of use can be the duration of play), information indicating whether the user with short silent periods jumps to the display interface of the target product by clicking on the display information, etc. The data types included in the first sample behavior data and the second sample behavior data can be consistent with the behavior data described in step 101 above.

[0077] Step 302: Use the first sample behavior data and the corresponding first identifier as input of a first initial value prediction model, use the first annotation information as the expected output of the initial value prediction model, and train the first initial value prediction model.

[0078] Step 303: Use the second sample behavior data and the corresponding second identifier as input of the initial value prediction model, use the second annotation information as the expected output of the initial value prediction model, and train the first initial value prediction model.

[0079] The first sample behavior data and the corresponding first identifier can be used as positive samples, and the second sample behavior data and the corresponding second identifier can be used as negative samples. The positive samples and negative samples are used to train the first initial value prediction model.

[0080] During training, the first sample behavior data can be input into the first initial value prediction model to obtain predicted data. The predicted data may include information such as predicted revenue, predicted usage duration, and predicted click probability for displayed information. Using a preset loss function (such as a mean square error loss function), the error between the first labeled information and the data actually predicted by the model is calculated to obtain a loss value representing the error. Using backpropagation and gradient descent, the parameters of the first initial value prediction model are adjusted to gradually reduce the loss value until convergence.

[0081] The method for training using the second sample behavior data and the second annotation information is the same as the above training method and will not be described in detail here.

[0082] Step 304: In response to the first initial value prediction model meeting the training end condition, the first initial value prediction model is determined as the first value prediction model.

[0083] Optionally, the training end conditions may include but are not limited to at least one of the following: the training duration exceeds a preset duration, the number of training times exceeds a preset number, and the loss value reaches a convergence condition.

[0084] It should be understood that the above steps 302-303 are training steps for a group of training samples. In actual training scenarios, multiple groups of training samples can be used to iteratively train the model, and the model after this iterative training is used as the initial model for the next iterative training, and finally the first value prediction model after training is obtained.

[0085] This embodiment collects first sample behavior data and second sample behavior data in advance for two scenarios: one with display information delivered and the other without display information delivered. The first sample behavior data and the second sample behavior data are used to train the first value prediction model. The first value prediction model can be trained specifically for short-cycle users in combination with actual scenarios. The trained first value prediction model has higher prediction accuracy and better scenario adaptability.

[0086] In some optional implementations, such as Figure 4 As shown, step 104 includes:

[0087] Step 1042: If the user type indicates that the target user is a long silent period user, extract a second value prediction model corresponding to the long silent period user.

[0088] The second value prediction model is pre-trained for users with long silent periods. The structure of the second value prediction model can be the same as that of the first value prediction model, but the training method is different. When training the second value prediction model, sample behavior data can be collected for users with long silent periods, and the model can be trained using the sample behavior data.

[0089] Step 105 includes:

[0090] Step 1055: Use the second value prediction model to predict the display value of the behavior data to obtain the display value score of the target user.

[0091] Specifically, the behavior data can be input into a second value prediction model to obtain a display value score.

[0092] Because users with long silent periods use the target product less frequently and are more sensitive to display information, they are more likely to click on, convert, and perform other actions on the target display information after viewing it. Therefore, there's no need to model the display value gain for users with long silent periods; predictions can be made directly using behavioral data. This allows the resulting display value score to more accurately match the target user's habits of using the target product, improving the accuracy of determining who receives the display information.

[0093] In some optional implementations, such as Figure 5 As shown, the second value prediction model is pre-trained according to the following steps:

[0094] Step 501: Acquire third sample behavior data recorded for a user with a long silence period.

[0095] The third sample behavior data has corresponding third annotation information. The third annotation information may include information about revenue earned from users with long silent periods, information about the duration of target product usage by users with long silent periods, information indicating whether users with long silent periods clicked on display information to jump to the target product's display interface, and the like. The data types included in the third sample behavior data may be consistent with the behavior data described in step 101 above.

[0096] Step 502: Use the third sample behavior data as input of a second initial value prediction model, use the third annotation information as the expected output of the second initial value prediction model, and train the second initial value prediction model.

[0097] During training, the third sample behavior data can be input into the second initial value prediction model to generate predicted data. This data can include information such as predicted revenue, predicted usage duration, and predicted click probability for displayed information. Using a preset loss function, the error between the third labeled information and the model's actual predicted data is calculated to obtain a loss value representing the error. Using backpropagation and gradient descent, the parameters of the second initial value prediction model are adjusted to gradually reduce the loss value until convergence.

[0098] Step 503: In response to the second initial value prediction model meeting the training end condition, the second initial value prediction model is determined as the second value prediction model.

[0099] The training end condition can be the same as above Figure 3 The training end conditions described in the corresponding embodiment are the same and will not be repeated here.

[0100] This embodiment targets long-term users and combines actual scenarios to specifically train the second value prediction model. The trained second value prediction model has higher prediction accuracy and better scenario adaptability.

[0101] In some optional implementations, such as Figure 6 As shown, the value prediction model includes a multi-task sharing network 601 and at least two value prediction networks 602.

[0102] like Figure 7 As shown, step 105 includes:

[0103] Step 1056: Use the multi-task shared network to extract features from the behavior data to obtain multi-task basic features.

[0104] The architecture of the value prediction model can be based on a multi-task model architecture such as MMoE. The multi-task shared network can include multiple layers of deep neural networks. The input behavioral data is calculated by the multi-task shared network, allowing multiple networks performing specific prediction tasks to share the extracted multi-task basic features.

[0105] Step 1057 : Use each value prediction network of the at least two value prediction networks to perform value prediction on the multi-task basic features to obtain at least two display value prediction values.

[0106] like Figure 6 As shown, at least two value prediction networks (including value prediction network 1 - value prediction network N, where N is a preset integer) respectively receive the input multi-task basic features and respectively perform corresponding prediction tasks (such as revenue prediction task, play duration prediction task, play frequency prediction task, etc.), and output display value prediction value 1 - display value prediction value N. When the value prediction model is implemented based on a multi-task model architecture such as MMoE, at least two value prediction networks are constructed in the form of task towers.

[0107] Step 1058: Fusing the at least two display value prediction values to obtain the display value score.

[0108] Specifically, a weight value corresponding to each display value prediction value can be set, and the weighted sum of each display value prediction value can be performed to obtain a display value score; or on the basis of the weighted sum, further calculation can be performed according to a preset numerical conversion strategy (such as normalization, etc.) to obtain a display value score.

[0109] The structure of the value prediction model provided in this embodiment can be applied to the first value prediction model and the second value prediction model in the above-mentioned embodiment, that is, the structures of the first value prediction model and the second value prediction model can be the same. When the target user is a user with a short silent period, the behavioral data needs to be merged with the first identifier or the second identifier and then input into the first value prediction model. When the target user is a user with a long silent period, the behavioral data can be directly input into the second value prediction model.

[0110] This embodiment achieves multi-dimensional prediction of behavioral data by setting a multi-task shared network and at least two value prediction networks in the value prediction model. The multiple display value prediction values obtained can more comprehensively represent the potential value of the target user, thereby improving the accuracy of the display value score prediction.

[0111] In some optional implementations, such as Figure 8 As shown, the at least two value prediction networks 602 include a usage time prediction network 6021 and at least one revenue prediction network 6022 (including Figure 8 The revenue prediction network shown is the revenue prediction network M, where M is a preset integer).

[0112] like Figure 9 As shown, step 1057 includes:

[0113] Step 10571: Utilize each revenue prediction network in the at least one revenue prediction network to perform revenue value prediction on the multi-task basic features respectively, and obtain at least one revenue prediction value as the display value prediction value corresponding to the at least one revenue prediction network.

[0114] The revenue forecast value represents the revenue that will be obtained from the target user by using the target display information within a first preset time period in the future (e.g., the next 7 days). For example, if the target display information is an advertisement, the revenue forecast value represents the estimated revenue that the advertiser will obtain from the target user by placing the advertisement.

[0115] Each of the revenue prediction networks described above can perform a task of predicting a specific type of revenue. For example, multiple revenue networks can be used to perform tasks such as total revenue prediction, membership revenue prediction, purchase revenue prediction, and on-demand revenue prediction.

[0116] Step 10572: Use the usage time prediction network to predict the usage time of the multi-task basic features to obtain the predicted usage time of the target user.

[0117] The predicted usage duration represents the duration of time that the target user uses the target website within a second preset duration in the future.

[0118] The second preset duration may be the same as or different from the first preset duration. The type of predicted usage duration may correspond to the actual type of the target display information. For example, if the target display information is information displaying a video playback website, the predicted usage duration may be the duration the target user accesses the video playback website and plays the video. For another example, if the target display information is information displaying a software download link, the predicted usage duration may be the duration the target user uses the software after downloading it.

[0119] like Figure 8 As shown, the output of each revenue forecast value and predicted usage time are integrated and calculated to obtain the display value score.

[0120] This embodiment sets up at least one revenue prediction network and a usage time prediction network to achieve the use of a value prediction model to predict the revenue and usage time brought to the target user. By using revenue and usage time, the value generated after the target display information is delivered to the target user can be more accurately represented, thereby further improving the accuracy of the model's value prediction.

[0121] Figure 10 A structural diagram of a display information push device based on user browsing frequency provided in an embodiment of the present application. Specifically comprising: an acquisition module 1001, for acquiring behavior data corresponding to a target user, wherein the behavior data includes browsing behavior data of the target user for a target product, and viewing data recorded by the target user performing related operations based on the display information related to the target product; a first determination module 1002, for determining the browsing frequency information of the target user based on the browsing behavior data; a second determination module 1003, for determining the user type to which the target user belongs based on the browsing frequency information; an extraction module 1004, for extracting a value prediction model corresponding to the user type; a prediction module 1005, for using the value prediction model to perform display value prediction on the behavior data to obtain a display value score of the target user, wherein the display value score represents the probability of the target user performing a preset operation after viewing the target display information; a sending module 1006, for sending the target display information to the terminal used by the target user if the display value score meets the preset display conditions.

[0122] In some optional implementations, the extraction module is further used to: if the user type indicates that the target user is a short silent period user, extract the first value prediction model corresponding to the short silent period user; the prediction module includes: an acquisition unit, used to obtain a preset first identifier indicating that display information is delivered to the target user, and a second identifier indicating that display information is not delivered to the target user; a merging unit, used to merge the first identifier and the behavior data to obtain first behavior data, and merge the second identifier and the behavior data to obtain second behavior data; a first prediction unit, used to use the first value prediction model to perform display value prediction on the first behavior data and the second behavior data respectively, to obtain a first display value score and a second display value score; a determination unit, used to determine the display value score of the target user based on the difference between the first display value score and the second display value score.

[0123] In some optional implementations, the first value prediction model is pre-trained according to the following steps: obtaining first sample behavior data recorded after display information is delivered to users with a short silent period, and second sample behavior data recorded without display information, wherein the first sample behavior data has corresponding first labeling information, and the second sample behavior data has corresponding first labeling information; using the first sample behavior data and the corresponding first identifier as the input of the first initial value prediction model, and the first labeling information as the expected output of the initial value prediction model, to train the first initial value prediction model; using the second sample behavior data and the corresponding second identifier as the input of the initial value prediction model, and the second labeling information as the expected output of the initial value prediction model, to train the first initial value prediction model; in response to the first initial value prediction model meeting the training end condition, determining the first initial value prediction model as the first value prediction model.

[0124] In some optional implementations, the extraction module is further used for: a first extraction unit, for extracting a second value prediction model corresponding to the long silence period user if the user type indicates that the target user is a long silence period user; the prediction module includes: a second prediction unit, for using the second value prediction model to perform display value prediction on the behavior data to obtain the display value score of the target user.

[0125] In some optional implementations, the second value prediction model is pre-trained according to the following steps: obtaining third sample behavior data recorded for users with long silence periods, wherein the third sample behavior data has corresponding third annotation information; using the third sample behavior data as the input of the second initial value prediction model, and using the third annotation information as the expected output of the second initial value prediction model, to train the second initial value prediction model; in response to the second initial value prediction model meeting the training end conditions, determining the second initial value prediction model as the second value prediction model.

[0126] In some optional implementations, the value prediction model includes a multi-task shared network and at least two value prediction networks; the prediction module includes: a second extraction unit, used to use the multi-task shared network to extract features from the behavior data to obtain multi-task basic features; a third prediction unit, used to use each value prediction network in the at least two value prediction networks to perform value prediction on the multi-task basic features to obtain at least two display value prediction values; and a fusion unit, used to fuse the at least two display value prediction values to obtain the display value score.

[0127] In some optional implementations, the at least two value prediction networks include a usage time prediction network and at least one revenue prediction network; the third prediction unit includes: a first prediction subunit, used to use each revenue prediction network in the at least one revenue prediction network to perform revenue value prediction on the multi-task basic features respectively, and obtain at least one revenue prediction value as the display value prediction value corresponding to the at least one revenue prediction network, wherein the revenue prediction value represents the size of the revenue obtained from the target user using the target display information within a first preset time in the future; a second prediction subunit, used to use the usage time prediction network to perform usage time prediction on the multi-task basic features, and obtain the predicted usage time of the target user, wherein the predicted usage time represents the time the target user uses the target website within a second preset time in the future.

[0128] The display information push device based on user browsing frequency provided in this embodiment can be as follows: Figure 10 The display information push device based on user browsing frequency shown in can execute all the steps of the above display information push methods based on user browsing frequency, thereby achieving the technical effects of the above display information push methods based on user browsing frequency. Please refer to the above related description for details. For the sake of brevity, it will not be repeated here.

[0129] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 11The electronic device 1100 shown includes: at least one processor 1101, a memory 1102, at least one network interface 1104 and another user interface 1103. The various components in the electronic device 1100 are coupled together via a bus system 1105. It is understood that the bus system 1105 is used to achieve connection and communication between these components. In addition to including a data bus, the bus system 1105 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, the bus system 1105 is not shown in FIG. Figure 11 Various buses are labeled as bus system 1105.

[0130] The user interface 1103 may include a display, a keyboard, or a pointing device (eg, a mouse, a trackball, a touchpad, or a touch screen).

[0131] It is understood that the memory 1102 in the embodiment of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDRSDRAM), enhanced synchronous DRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus random access memory (DRRAM). The memory 1102 described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0132] In some embodiments, the memory 1102 stores the following elements, executable units, or data structures, or a subset thereof, or an extended set thereof: an operating system 11021 and application programs 11022 .

[0133] Among them, the operating system 11021 includes various system programs, such as the framework layer, core library layer, and driver layer, which are used to implement various basic services and handle hardware-based tasks. Application programs 11022 include various application programs, such as media players and browsers, which are used to implement various application services. Programs that implement the methods of the embodiments of the present application can be included in application programs 11022.

[0134] In this embodiment, by calling the program or instructions stored in the memory 1102, specifically, the program or instructions stored in the application 11022, the processor 1101 is configured to execute the method steps provided by each method embodiment, for example, including:

[0135] Obtaining behavioral data corresponding to the target user, wherein the behavioral data includes the target user's browsing behavior data for the target product, and viewing data recorded when the target user performs related operations based on display information related to the target product; determining the target user's browsing frequency information based on the browsing behavior data; determining the user type to which the target user belongs based on the browsing frequency information; extracting a value prediction model corresponding to the user type; using the value prediction model, performing display value prediction on the behavioral data to obtain the target user's display value score, wherein the display value score represents the probability of the target user performing a preset operation after viewing the target display information; if the display value score meets the preset display conditions, sending the target display information to the terminal used by the target user.

[0136] The methods disclosed in the above embodiments of the present application can be applied to or implemented by processor 1101. Processor 1101 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in processor 1101 or by software instructions. The above processor 1101 may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The methods, steps, and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of the present application can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software units in the decoding processor. The software units can be located in a storage medium mature in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory 1102 , and the processor 1101 reads the information in the memory 1102 and completes the steps of the above method in combination with its hardware.

[0137] It is understood that the embodiments described herein may be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit may be implemented in 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), general-purpose processors, controllers, microcontrollers, microprocessors, or other electronic units or combinations thereof for performing the above-mentioned functions of the present application.

[0138] For software implementation, the techniques described above can be implemented by a unit that performs the functions described above. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or external to the processor.

[0139] The electronic device provided in this embodiment may be Figure 11 The electronic device shown in can execute all the steps of the above-mentioned methods for pushing display information based on user browsing frequency, thereby achieving the technical effects of the above-mentioned methods for pushing display information based on user browsing frequency. For details, please refer to the above related description. For the sake of brevity, it will not be repeated here.

[0140] The present application also provides a storage medium (computer-readable storage medium). The storage medium stores one or more programs. The storage medium may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as read-only memory, flash memory, hard disk, or solid-state drive; and the memory may also include a combination of the aforementioned types of memory.

[0141] When one or more programs in the storage medium can be executed by one or more processors, the above-mentioned method for pushing display information based on user browsing frequency executed on the electronic device side is implemented.

[0142] The processor is configured to execute a program stored in the memory to implement the following steps of a method for pushing display information based on user browsing frequency, which is performed on the electronic device side:

[0143] Obtaining behavioral data corresponding to the target user, wherein the behavioral data includes the target user's browsing behavior data for the target product, and viewing data recorded when the target user performs related operations based on display information related to the target product; determining the target user's browsing frequency information based on the browsing behavior data; determining the user type to which the target user belongs based on the browsing frequency information; extracting a value prediction model corresponding to the user type; using the value prediction model, performing display value prediction on the behavioral data to obtain the target user's display value score, wherein the display value score represents the probability of the target user performing a preset operation after viewing the target display information; if the display value score meets the preset display conditions, sending the target display information to the terminal used by the target user.

[0144] Professionals should also be further aware that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different circuits to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0145] The steps of the circuits or algorithms described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0146] It should be understood that the terms used herein are for the purpose of describing specific example embodiments only and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "said" as used herein may also be meant to include plural forms. The terms "comprise", "include", "contain" and "have" are inclusive and therefore specify the presence of stated features, steps, operations, elements and / or parts, but do not exclude the presence or addition of one or more other features, steps, operations, elements, parts, and / or combinations thereof. The steps, processes, and operations described herein are not to be construed as necessarily requiring them to be performed in the specific order described or illustrated, unless the order of execution is clearly indicated. It should also be understood that additional or alternative steps may be used.

[0147] The foregoing is merely a list of specific embodiments of the present application, intended to enable those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the broadest scope consistent with the principles and novel features of the present application.

Claims

1. A method for pushing display information based on user browsing frequency, characterized in that: The method comprises: Obtaining behavior data corresponding to the target user, wherein the behavior data includes browsing behavior data of the target user for the target product, and viewing data recorded when the target user performs related operations based on display information related to the target product; Determining browsing frequency information of the target user based on the browsing behavior data; Determining the user type of the target user based on the browsing frequency information; Extracting a value prediction model corresponding to the user type; Using the value prediction model, performing display value prediction on the behavior data to obtain a display value score of the target user, wherein the display value score represents the probability of the target user performing a preset operation after viewing the target display information; If the display value score meets the preset display conditions, the target display information is sent to the terminal used by the target user.

2. The method according to claim 1, characterized in that The extracting the value prediction model corresponding to the user type includes: If the user type indicates that the target user is a short silent period user, extracting a first value prediction model corresponding to the short silent period user; The step of using the value prediction model to predict the display value of the behavior data to obtain the display value score of the target user includes: Obtaining a preset first flag indicating that the display information is delivered to the target user, and a second flag indicating that the display information is not delivered to the target user; Merging the first identifier and the behavior data to obtain first behavior data, and merging the second identifier and the behavior data to obtain second behavior data; Using the first value prediction model, respectively predict the display value of the first behavior data and the second behavior data to obtain a first display value score and a second display value score; An exhibition value score of the target user is determined based on a difference between the first exhibition value score and the second exhibition value score.

3. The method according to claim 2, characterized in that The first value prediction model is pre-trained according to the following steps: Obtaining first sample behavior data recorded after display information is delivered to users with a short silent period, and second sample behavior data recorded without display information being delivered, wherein the first sample behavior data has corresponding first labeling information, and the second sample behavior data has corresponding first labeling information; Using the first sample behavior data and the corresponding first identifier as input to a first initial value prediction model, and using the first annotation information as the expected output of the initial value prediction model, to train the first initial value prediction model; Using the second sample behavior data and the corresponding second identifier as input to the initial value prediction model, and using the second annotation information as the expected output of the initial value prediction model, to train the first initial value prediction model; In response to the first initial value prediction model meeting the training end condition, the first initial value prediction model is determined as the first value prediction model.

4. The method according to claim 1, wherein The extracting the value prediction model corresponding to the user type includes: If the user type indicates that the target user is a long silent period user, extracting a second value prediction model corresponding to the long silent period user; The step of using the value prediction model to predict the display value of the behavior data to obtain the display value score of the target user includes: The second value prediction model is used to predict the display value of the behavior data to obtain the display value score of the target user.

5. The method according to claim 4, characterized in that The second value prediction model is pre-trained according to the following steps: Acquire third sample behavior data recorded for a user with a long silence period, wherein the third sample behavior data has corresponding third annotation information; Using the third sample behavior data as input of a second initial value prediction model and the third labeled information as expected output of the second initial value prediction model, training the second initial value prediction model; In response to the second initial value prediction model meeting the training end condition, the second initial value prediction model is determined as the second value prediction model.

6. The method according to any one of claims 1 to 5, characterized in that The value prediction model includes a multi-task shared network and at least two value prediction networks; The step of using the value prediction model to predict the display value of the behavior data to obtain the display value score of the target user includes: Using the multi-task shared network, extracting features from the behavior data to obtain multi-task basic features; Using each of the at least two value prediction networks, perform value prediction on the multi-task basic features to obtain at least two display value prediction values; The at least two display value prediction values are fused to obtain the display value score.

7. The method according to claim 6, characterized in that The at least two value prediction networks include a usage duration prediction network and at least one revenue prediction network; The method of using each of the at least two value prediction networks to perform value prediction on the multi-task basic features to obtain at least two display value prediction values includes: Using each of the at least one revenue prediction networks, respectively predict revenue values for the multi-task basic features, and obtain at least one revenue prediction value as a display value prediction value corresponding to the at least one revenue prediction network, wherein the revenue prediction value represents the amount of revenue that can be obtained from the target user by using the target display information within a first preset time period in the future; The usage time prediction network is used to predict the usage time of the multi-task basic features to obtain the predicted usage time of the target user, wherein the predicted usage time represents the time the target user uses the target website within a second preset time period in the future.

8. A device for pushing display information based on user browsing frequency, characterized in that: The device comprises: an acquisition module, configured to acquire behavior data corresponding to a target user, wherein the behavior data includes browsing behavior data of the target user for a target product, and viewing data recorded when the target user performs related operations based on display information related to the target product; A first determining module is configured to determine browsing frequency information of the target user based on the browsing behavior data; A second determining module is configured to determine the user type to which the target user belongs based on the browsing frequency information; An extraction module, configured to extract a value prediction model corresponding to the user type; A prediction module, configured to use the value prediction model to perform display value prediction on the behavior data to obtain a display value score of the target user, wherein the display value score represents the probability of the target user performing a preset operation after viewing the target display information; The sending module is configured to send the target display information to a terminal used by the target user if the display value score meets a preset display condition.

9. An electronic device, characterized in that: include: Memory for storing computer programs; The processor is configured to execute the computer program stored in the memory, and when the computer program is executed, implements the method for pushing display information based on user browsing frequency as described in any one of claims 1 to 7 above.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for pushing display information based on user browsing frequency described in any one of claims 1 to 7 is implemented.

Images