CN114764472A

CN114764472A - Content pushing method and device

Info

Publication number: CN114764472A
Application number: CN202110042030.9A
Authority: CN
Inventors: 林岳
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2021-01-13
Filing date: 2021-01-13
Publication date: 2022-07-19

Abstract

The application discloses a content push method and a content push device; the method comprises the steps of obtaining a full amount of objects of a content pushing task; extracting the full-scale objects to obtain a plurality of experimental objects; pushing task data to the experimental object based on the content pushing task to obtain feedback information of the experimental object to the task data, wherein the task data is associated with the content pushing task; determining a data label of the experimental object according to the feedback information; training the prediction model based on the experimental object and the data label to obtain a trained prediction model; determining a target object of a content pushing task from the full-scale objects through the trained prediction model so as to push the content through the target object; the content pushing efficiency can be improved.

Description

Content pushing method and device

Technical Field

The application relates to the field of internet, in particular to a content pushing method and device.

Background

In the internet era, besides information is obtained by means of active search and the like, a large amount of contents actively pushed by a client exist, for example, advertisements on a webpage and the like, when the advertisements are pushed, in order to improve the return rate of the advertisements, a pushing object can be selected.

In the process of research and practice of the prior art, the inventor of the present application finds that the prior art needs to consume more manpower, resulting in low content push efficiency.

Disclosure of Invention

The embodiment of the application provides a content pushing method and device, which can improve content pushing efficiency.

The embodiment of the application provides a content pushing method, which comprises the following steps:

acquiring a full object of a content pushing task;

extracting the full-scale objects to obtain a plurality of experimental objects;

pushing task data to the experimental object based on the content pushing task to acquire feedback information of the experimental object to the task data, wherein the task data is associated with the content pushing task;

determining a data label of the experimental object according to the feedback information;

training a prediction model based on the experimental object and the data label to obtain a trained prediction model;

and determining a target object of the content pushing task from the full-scale objects through the trained prediction model so as to push the content through the target object.

Accordingly, the present application provides a content push apparatus, comprising:

the acquisition module is used for acquiring a full amount of objects of the content pushing task;

The extraction module is used for extracting the full-scale objects to obtain a plurality of experimental objects;

the pushing module is used for pushing task data to the experimental object based on the content pushing task so as to obtain feedback information of the experimental object to the task data, and the task data is associated with the content pushing task;

the label determining module is used for determining the data label of the experimental object according to the feedback information;

the training module is used for training a prediction model based on the experimental object and the data label to obtain a trained prediction model;

and the object determining module is used for determining a target object of the content pushing task from the full-scale objects through the trained prediction model so as to push the content through the target object.

In some embodiments, the feedback information includes positive feedback information and negative feedback information, the data tags include positive tags and negative tags, the tag determination module includes a first determination sub-module and a second determination sub-module, wherein,

the first determining submodule is used for determining that the data label of the experimental object is a forward label when the feedback information of the experimental object is forward feedback information;

And the second determining submodule is used for determining that the data label of the experimental object is a negative label when the feedback information of the experimental object is negative feedback information.

In some embodiments, the content push apparatus further comprises:

the reference module is used for extracting the full-scale objects to obtain a plurality of reference objects;

the history module is used for acquiring the history feedback information of the experimental object;

the change module is used for determining reference feedback change information of the reference object, wherein the reference feedback change information comprises the change degree of the feedback information of the reference object in a preset time period;

at this time, the tag determination module includes a tag determination sub-module, wherein:

and the label determining submodule is used for determining the data label of the experimental object based on the historical feedback information, the reference feedback change information and the feedback information.

In some embodiments, the tag determination submodule includes a calculation unit and a determination unit, wherein,

the calculation unit is used for calculating the experimental feedback change information of the experimental object according to the historical feedback information, the reference feedback change information and the feedback information; (ii) a

And the determining unit is used for determining the data label of the experimental object according to the experimental feedback change information of the experimental object.

In some embodiments, the data tag includes a positive tag and a negative tag, and the determining unit is specifically configured to:

when the experiment feedback change information of the experimental object is larger than a preset threshold value, determining that the data label of the experimental object is a forward label;

and when the experiment feedback change information of the experimental object is not larger than the preset threshold value, determining that the data label of the experimental object is a negative label.

In some embodiments, the preset threshold includes a preset first threshold and a preset second threshold, the data tag includes a positive tag, a negative tag and a neutral tag, and the determining unit is specifically configured to:

when the experiment feedback change information of the experiment object is larger than or equal to a preset first threshold value, determining that the data label of the experiment object is a forward label;

when the experiment feedback change information of the experimental object is smaller than the preset first threshold and larger than the preset second threshold, determining that the data label of the experimental object is a neutral label;

and when the experiment feedback change information of the experiment object is smaller than or equal to the preset second threshold value, determining that the data label of the experiment object is a negative label.

In some embodiments, the object determination module includes a construction sub-module, an input sub-module, and a determination sub-module, wherein,

the construction submodule is used for carrying out feature construction on the full-scale object to obtain a full-scale feature vector;

the input submodule is used for inputting the full-scale feature vector into the trained prediction model to obtain label prediction information and confidence coefficient of the full-scale object;

and the determining submodule is used for determining a target object of the content pushing task from the full-scale objects based on the label prediction information and the confidence coefficient thereof so as to push the content through the target object.

In some embodiments, the determination submodule is specifically configured to:

when the prediction tag information of the full-scale object is a target data tag, determining the full-scale object as a candidate object;

based on the confidence of each candidate object, a target object is determined from all candidate objects.

In some embodiments, the training module is specifically configured to:

performing feature construction on the experimental object to obtain an experimental feature vector;

obtaining a model prediction result of the experimental object through the experimental feature vector and a prediction model;

and training the prediction model based on the model prediction result and the data label to obtain the trained prediction model.

Correspondingly, the embodiment of the present application further provides a storage medium, where the storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute any one of the content push methods provided in the embodiment of the present application.

Accordingly, embodiments of the present application further provide a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements any one of the content push methods provided in the embodiments of the present application when executing the computer program.

The method and the device can obtain the full objects of the content pushing task; extracting the full-scale objects to obtain a plurality of experimental objects; pushing task data to the experimental object based on the content pushing task to obtain feedback information of the experimental object to the task data, wherein the task data is associated with the content pushing task; determining a data label of the experimental object according to the feedback information; training the prediction model based on the experimental object and the data label to obtain a trained prediction model; and determining a target object of the content push task from the full-scale objects through the trained prediction model so as to push the content through the target object.

According to the method and the device, the experimental object can be extracted from the full-scale object firstly, content pushing is carried out on the experimental object, the data label of the experimental object is determined according to feedback information of the experimental object on pushed task data, then the prediction model is trained according to the experimental object and the data label of the experimental object, and finally the target pushing object is determined from the full-scale data through the trained prediction model to carry out content pushing.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic view of a scenario of a content push system provided in an embodiment of the present application;

fig. 2 is a schematic flowchart of a content push method provided in an embodiment of the present application;

fig. 3 is another schematic flow chart of a content push method provided by an embodiment of the present application;

Fig. 4 is a schematic structural diagram of a content pushing apparatus provided in an embodiment of the present application;

fig. 5 is a schematic structural diagram of a computer device provided in an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application, and it is obvious that the embodiments described in the present application are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

The content push system may be integrated in a computer device, the computer device may include at least one of a terminal and a server, and the server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing a cloud computing service. The terminal may be, but is not limited to, a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and the like. The terminal and the server may be directly or indirectly connected through wired or wireless communication, and the application is not limited herein.

Referring to fig. 1, the content push system may be integrated on a computer device such as a terminal or a server, and the computer device may obtain a full object of a content push task; extracting the full-scale objects to obtain a plurality of experimental objects; pushing task data to the experimental object based on the content pushing task to obtain feedback information of the experimental object to the task data, wherein the task data is associated with the content pushing task; determining a data label of the experimental object according to the feedback information; training the prediction model based on the experimental object and the data label to obtain a trained prediction model; and determining a target object of the content push task from the full-scale objects through the trained prediction model so as to push the content through the target object.

It should be noted that the scenario diagram of the content push system shown in fig. 1 is merely an example, and the content push system and the scenario described in the embodiment of the present application are for more clearly illustrating the technical solution of the embodiment of the present application, and do not form a limitation to the technical solution provided in the embodiment of the present application, and it is known by a person of ordinary skill in the art that the technical solution provided in the embodiment of the present application is also applicable to similar technical problems with the evolution of a content push device and the occurrence of a new service scenario.

The following are detailed descriptions. In this embodiment, a content push method will be described in detail, where the content push method may be integrated on a computer device, as shown in fig. 2, and fig. 2 is a schematic flowchart of the content push method provided in this embodiment of the present application. The content pushing method can comprise the following steps:

101. and acquiring a full amount of objects of the content pushing task.

The full-scale object comprises all potential objects which can be used for content pushing, the full-scale object generally comprises a real user, and under the conditions of testing, experiments and the like, the full-scale object also comprises a virtual object which simulates a user, such as a test script and the like.

The full-amount object may be distinguished by a full-amount object identifier that may serve as a unique identifier, such as a character, a two-dimensional code, and a text, for example, registered account information in the client, an Internet Protocol Address (IP) of a computer accessing a web page, and the like.

In the present application, the content pushing task includes content pushing to a target object determined from a full amount of objects, the content pushing may include various types, such as advertisement pushing, news pushing, activity information pushing, promotion pushing, play recommendation pushing, and the like, the content pushing may be performed in the form of pop-up windows, messages, gaps directly displayed in the content viewed by the user, and the like, the purpose of the content pushing is to enable the object to actively view the pushed content, however, in an actual application scenario, not every object receiving the pushed content will actively check, but may also directly ignore, at this time, the content pushing becomes an invalid behavior, in this case, the purpose of the content push task is to enable the object to view the probability of the content, and by selecting a target object with higher interest and higher viewing possibility for the pushed content, the purpose of improving the content viewing probability is achieved.

Specifically, the manner of obtaining the full amount object may include multiple manners, for example, the full amount object may be obtained directly from a computer device that integrates the solution of the present application, or for example, a request (such as a server or the like) may be sent to a computer device that stores the full amount data, the full amount object returned by the computer device according to the request is received, and the like.

For example, the content push task may be to push an advertisement on a personal social page of a user using instant messaging software, and account ids of all accounts using a personal social page function in the instant messaging software may be obtained (i.e., a full amount object is obtained).

102. And extracting the total amount of the objects to obtain a plurality of experimental objects.

The experimental object may include a part of the full-scale object, specifically, the full-scale object is extracted, the manner of obtaining the experimental object may include multiple manners, for example, random extraction may be performed, for example, the full-scale object may be grouped according to the property of the full-scale object, extraction may be performed in each group, finally, the extraction results of each group are integrated together to obtain the experimental object, for example, the full-scale object may be sampled for multiple times in a period of time, and the results of each sampling are integrated together to obtain the experimental object, and so on.

For example, the total number of objects may be divided into a plurality of groups according to the age information carried by the account id, random extraction may be performed in each group, and finally, the extraction results obtained from all the groups may be integrated to obtain the experimental object.

103. And pushing task data to the experimental object based on the content pushing task to acquire feedback information of the experimental object to the task data, wherein the task data is associated with the content pushing task.

The task data may include data of content that the content push task wants to push, for example, when the content push task is an advertisement push task, the task data may include an advertisement to be pushed, for example, when the content push task is a voting push task, the task data may include voting information, and the task data may further include questionnaire survey content, promotional content, and the like.

The feedback information may include reaction information of the experimental object based on the task data, and the feedback information may include a duration of the object staying on a page where the task data is located, a frequency of operations pointed by content of the task data, and the like, for example, a frequency of the object using a service corresponding to the advertisement content, a duration of the object staying on a page where the advertisement content is located, and the like.

The feedback information may include responses of the experimenter to the task data, such as clicking advertisements, getting coupons, purchasing goods, watching videos, filling in questionnaires, clicking links, clicking messages and the like, and the feedback information may also include the invisibility of the experimenter to the task data, that is, when the experimenter does not respond to the task data at all, the feedback information is recorded as a feedback information.

Specifically, task data corresponding to the content pushing task may be determined first, then task data pushing may be performed on each experimental object, and feedback information of the experimental object on the task data is determined according to a reaction of the experimental object on content displayed on a page based on the task data or a reaction of the experimental object on the page including the content within a period of time.

For example, task data (advertisement data) corresponding to a content push task (advertisement task) is determined, the advertisement data is pushed to the experimental object, and feedback information of the experimental object to the advertisement data is obtained according to the reaction of the experimental object.

104. And determining the data label of the experimental object according to the feedback information.

The data label can include a representation of feedback information of the experimental object, the data label can include a character form which is easy to distinguish and read, different data labels refer to different feedback information, specifically, the process of determining the data label of the experimental object according to the feedback information can be flexibly performed along with the difference of the feedback information, for example, the feedback information can be a continuous integer and can be set into an interval, the data label of the experimental object corresponding to the feedback information is determined according to the interval where the feedback information is located, for example, the feedback information can be a limited number of simple texts, at the moment, one data label can be corresponding to each text, and then the data label of each experimental object is determined.

In some embodiments, the feedback information includes positive feedback information and negative feedback information, the data tag includes a positive tag and a negative tag, and the step of "extracting the full-scale subject to obtain the plurality of experimental subjects" may include:

when the feedback information of the experimental object is the forward feedback information, determining that the data label of the experimental object is a forward label; and when the feedback information of the experimental object is negative feedback information, determining that the data label of the experimental object is a negative label.

For example, when the content push task is an advertisement push task, the feedback information of the experimental object may be whether to click an advertisement, if an experimental object clicks the advertisement, the feedback information of the experimental object is positive feedback information, if an experimental object does not click the advertisement, the feedback information of the experimental object is negative feedback information, then, the data tag of the experimental object may be determined according to the feedback information of the experimental object, and if the feedback information of the experimental object is positive feedback information, the data tag of the experimental object is a positive tag; and when the feedback information of an experimental object is negative feedback information, determining that the data label of the experimental object is a negative label.

In some embodiments, the content push method further comprises:

extracting the full-scale objects to obtain a plurality of reference objects; acquiring historical feedback information of an experimental object; determining reference feedback change information of the reference object, wherein the reference feedback change information comprises the change degree of the feedback information of the reference object in a preset time period;

at this time, the step of "determining the data label of the experimental subject according to the feedback information" may include:

and determining the data label of the experimental object based on the historical feedback information, the reference feedback change information and the feedback information.

In some embodiments, determining the data label of the subject needs to be performed based on historical feedback information and feedback information of the subject and reference feedback change information of the reference subject,

the reference object may include a part of the full-scale object, the reference object may correspond to the experimental object, and similar to the experimental group and the control group, the reference object may exclude interference of factors other than content push on determination of the data tag to some extent.

The extraction mode of the reference object may be similar to that of the experimental object, and is not described herein again, when the reference object and the experimental object are extracted, the reference object and the experimental object are extracted according to a certain proportion from the full-scale object, where the proportion may be determined according to the number of the full-scale object, the number of the experimental object and the reference object, and other factors, for example, the proportion may be 1%, and in addition, the number of the reference object and the experimental object may be flexibly determined based on the training requirement of the prediction model and other factors, for example, the number of the reference object and the experimental object may be set to 10 thousands.

For example, if the content push is to push an advertisement on an information flow page (the information flow may include a video, a text, an image, etc.), the advertisement may be pushed to the information flow page of the experimental object within a preset time period, the average staying time (i.e., feedback information) of the object on the information flow page after the advertisement is pushed is recorded, the advertisement push is not performed on the information flow page of the reference object, the average staying time of the reference object on the information flow page at the start stage and the end stage of the preset time period are respectively recorded, and then the reference feedback change information of the reference object is obtained.

The reference feedback change information includes a change degree of the feedback information of the reference object in a preset time period, the calculation mode of the reference feedback change information may include multiple modes, for example, the average stay time of the termination stage is directly divided by the average stay time of the start stage, for example, the content push task may push the promotion information of the shopping website, the reference feedback change information may be that the average number of orders of the termination stage object at a preset time end is different from the average number of orders of the start stage object, the calculation mode and the expression mode of the reference feedback change information may also change in different content push tasks, and the calculation mode and the expression mode of the reference feedback change information may be flexibly determined according to the actual situation in the application scene without limitation.

The historical feedback information may include reactions of the subjects before pushing the task data, for example, the historical feedback information may include average stay time of the subjects on the information flow page before pushing the advertisement to the information flow page, and for example, the historical feedback information may include average number of orders of the subjects before pushing the promotion information, and the like.

The process of determining the data tag according to the historical feedback information and the feedback information of the experimental object and the reference feedback change information of the reference object may include various manners, for example, may be performed in a fusion manner such as weighted average, difference calculation, summation, or may also be performed in a ranking manner, a comparison manner, or the like.

In some embodiments, the step of "determining the data label of the subject based on the historical feedback information, the reference feedback change information, and the feedback information" may comprise:

calculating experimental feedback change information of the experimental object according to the historical feedback information, the feedback information and the reference feedback change information; and determining the data label of the experimental object through the experimental feedback change information.

Wherein, the experimental feedback change information may include a change degree of feedback information of the experimental object within a preset time period, and the preset time period may be a period during which the task is pushed, specifically, a data tag of the experimental object is determined, the initial feedback change information of the experimental object may be first calculated, a calculation manner of the initial feedback change information is similar to that of the reference feedback change information, for example, the feedback information may be divided by the historical feedback information, or the historical feedback information may be subtracted by the feedback information, and the like, specifically, the initial feedback change information and the reference feedback change information may be flexibly set in an actual application process, and then, the experimental feedback change information of the experimental object may be determined based on the initial feedback change information and the reference feedback change information, for example, the initial feedback change information and the reference feedback change information of the experimental object may be compared, and the experimental feedback change information may be obtained according to the comparison result, for another example, the initial feedback change information and the reference feedback change information may be weighted and summed to obtain the experimental feedback change information, and the like.

And finally, determining the data label of the experimental object through the experimental feedback change information, for example, determining the data label corresponding to the experimental feedback change information of the experimental object according to the mapping relationship.

For example, the initial feedback change information C of the experimental object 11 may be obtained according to the historical feedback information L and the feedback information F, the initial feedback change information C of the experimental object 11 is compared with the reference feedback change information K, the experimental feedback change information S of the experimental object 11 is determined according to the comparison result, and the data tag of the experimental object 11 is determined to be the data tag 1.

In some embodiments, the step of determining the data label of the subject according to the experimental feedback variation information of the subject may include:

when the experiment feedback change information of the experimental object is larger than a preset threshold value, determining that the data label of the experimental object is a forward label; and when the experiment feedback change information of the experimental object is not larger than the preset threshold value, determining that the data label of the experimental object is a negative label.

For example, the experiment feedback change information of the experimental object obtained through calculation may be a real number, when the data tag is determined, a magnitude relationship between the experiment feedback change information of the experimental object and a preset threshold may be compared, and the data tag of the experimental object is obtained according to a comparison result, where the preset threshold may be 0, if the experiment feedback change information of an experimental object is greater than 0, the data tag of the experimental object may be determined to be a positive tag, if the experiment feedback change information of an experimental object is less than or equal to 0, the data tag of the experimental object may be determined to be a negative tag, and the positive tag and the negative tag may be identified by a character, where the positive tag may be identified as 1 and the negative tag may be identified as 0.

In some embodiments, the preset threshold includes a preset first threshold and a preset second threshold, the data labels include positive and negative labels and a neutral label, and the step of determining the data label of the experimental subject according to the experimental feedback change information of the experimental subject may include:

when the experiment feedback change information of the experimental object is larger than or equal to a preset first threshold value, determining that the data label of the experimental object is a forward label; when the experiment feedback change information of the experiment object is smaller than a preset first threshold and larger than a preset second threshold, determining that the data label of the experiment object is a neutral label; and when the experiment feedback change information of the experiment object is less than or equal to a preset second threshold value, determining that the data label of the experiment object is a negative label.

For example, the experiment feedback change information of the experimental object may be a real number, when the data tag is determined, the magnitude relationship between the experiment feedback change information of the experimental object and a preset threshold may be compared, and the data tag of the experimental object is obtained according to the comparison result, where the preset threshold may include a plurality of data tags, and the data tag may also include a plurality of data tags, for example, the preset threshold may include a preset first threshold and a preset second threshold, and the data tag may include a positive tag, a negative tag and a neutral tag, and if the preset first threshold and the preset second threshold are 1 and-1, respectively, when the experiment feedback change information of the experimental object is greater than or equal to 1, the data tag of the experimental object may be determined to be the positive tag; when the experiment feedback change information of the experimental object is less than 1 and greater than-1, determining that the data label of the experimental object is a neutral label; when the experimental feedback change information of the experimental subject is less than or equal to-1, the data label of the experimental subject can be determined to be a negative label.

105. And training the prediction model based on the experimental object and the data label to obtain the trained prediction model.

The prediction model may include a model capable of predicting a reaction of the object to the content push, the prediction model may include a classification model, and the common classification model may include: the method includes a tree model, a linear regression model, a Gradient lifting model (such as an eXtreme Gradient lifting model (XGBoost)), a random forest model, a Long-Short Term Memory network model (LSTM), a neural network, and the like, and specifically, the method can be flexibly selected in a practical process without limitation.

For example, the prediction model may be a gradient lifting model, and the constructed prediction model may be trained according to the experimental object and the data label thereof, so as to obtain a trained prediction model a.

In some embodiments, the step of training the prediction model based on the experimental subject and the data label to obtain the trained prediction model may include:

carrying out feature construction on an experimental object to obtain an experimental feature vector; obtaining a model prediction result of an experimental object through the experimental characteristic vector and a prediction model; and training the prediction model based on the model prediction result and the data label to obtain the trained prediction model.

Specifically, when the Feature is constructed, the Feature construction can be flexibly performed according to the experimental object identifier, the content prediction task, and the like of the experimental object, for example, common processes of the Feature construction may include Binning (Binning), One-Hot Encoding (One-Hot Encoding), Feature Hashing (Hashing), nesting (Embedding), logarithm (Log Transformation), Feature Scaling (Scaling), Normalization (Normalization), Feature Interaction (Feature Interaction), and the like.

The method includes inputting an experiment feature vector into a prediction model to obtain a model prediction result of an experimental object, calculating a loss value of the training according to a data tag of the experimental object and the model prediction result, adjusting parameters of the prediction model, and implementing a training process to obtain the trained prediction model, and specifically, an algorithm referred by the adjusted parameters may include various algorithms, such as random Gradient Descent (SGD), random Gradient Descent using Momentum (momentd), Adaptive Gradient (Adaptive Gradient), and the like.

106. And determining a target object of the content push task from the full-scale objects through the trained prediction model so as to push the content through the target object.

For example, the total objects except the determined experimental object may be target total objects, the target total objects are sequentially input into the trained prediction model to obtain output information of each target total object, the target objects are obtained by screening from the target total objects according to the output information, and content pushing is performed on the target objects.

In some embodiments, the step of determining, by the trained predictive model, a target object of the content push task from the full volume objects to push the content through the target object may include:

performing feature construction on the full-scale object to obtain a full-scale feature vector; inputting the full quantity feature vector into the trained prediction model to obtain label prediction information and confidence coefficient of the full quantity object; and determining a target object of the content push task from the full-scale objects based on the label prediction information and the confidence coefficient thereof so as to push the content through the target object.

Before inputting the full-scale object into the trained prediction model, feature construction needs to be performed on the full-scale object, the feature construction mode and steps are similar to those of the experimental object, and are not repeated here, and after the feature construction is performed on the full-scale object to obtain a full-scale feature vector, the full-scale feature vector can be input into the trained prediction model to obtain label prediction information and confidence of the full-scale object.

The tag prediction information may include prediction of a data tag of the global object, where the data tag is determined by the feedback information, and the tag prediction information/data tag represents performance of a corresponding full-scale object in a content push task (e.g., advertisement click, increase in purchase times, etc.). The confidence level may include a confidence level of the tag prediction information, and the higher the confidence level is, the higher the confidence level of the tag prediction information is.

In some embodiments, the step of "determining a target object for the content push task from the full objects based on the tag prediction information and its confidence" may include:

when the predicted tag information of the full-scale object is a target data tag, determining the full-scale object as a candidate object; based on the confidence of each candidate object, a target object is determined from all candidate objects.

For example, the target data tag may be data tag 1, the full number of objects with data tags of 1 are reserved as candidate objects, ranking is performed according to the confidence of the tag prediction information of each candidate object, and then the target object is determined from the candidate objects according to the ranking information.

The method described in the above embodiments is further illustrated in detail by way of example.

The content push method will be described with reference to a content push system integrated in a computer device as an example, as shown in fig. 3, fig. 3 is a schematic flow diagram of the content push method provided in the embodiment of the present application. The content pushing method can comprise the following steps:

201. the computer device obtains a full object for the content push task.

For example, the content push task may be an e-commerce advertisement push task, the full-volume object may be all users who have used the e-commerce advertisement, and the computer device may read the full-volume users of the e-commerce advertisement push task from the database.

202. And extracting the full-scale objects by the computer equipment to obtain a plurality of experimental objects.

For example, the computer device may randomly draw 1% of all users who have used the e-commerce as experimental subjects.

203. The computer equipment extracts the full-scale objects to obtain a plurality of reference objects.

For example, the computer device may randomly draw 1% of all users who have used the e-commerce as a reference object.

204. The computer device pushes task data to the experimental object based on the content pushing task to obtain feedback information of the experimental object, wherein the task data is associated with the content pushing task.

For example, the computer device may push an e-commerce advertisement (i.e., task data) to the experimental object, the e-commerce advertisement may be presented to the experimental object in a form of, for example, a character, an image, an animation, a video, and the like, and specifically, the presentation manner may include displaying the e-commerce advertisement and the like on a page frequently browsed by the experimental object, so as to obtain feedback information of the experimental object, where the feedback information may be a purchase frequency of the experimental object in the last week of the e-commerce after the e-commerce advertisement is pushed to the experimental object, and the e-commerce advertisement is determined according to an e-commerce advertisement pushing task.

205. The computer equipment acquires the historical feedback information of the experimental object and determines the reference feedback change information of the reference object.

For example, before the computer device delivers the e-commerce advertisement to the experimental subject, the computer device obtains historical feedback information of the experimental subject, where the historical feedback information includes an average weekly purchase frequency (an average of the weekly purchase frequency) of the experimental subject on the e-commerce.

The reference feedback change information may reflect a change degree of the feedback information of the reference object during the period that the reference object pushes the e-commerce advertisement to the experimental object, for example, before the computer device puts the e-commerce advertisement to the experimental object and after the computer device puts the e-commerce advertisement to the experimental object, the weekly average purchase frequency 1 and the weekly average purchase frequency 2 of the reference object are respectively obtained, the weekly average purchase frequency 1 is subtracted from the weekly average purchase frequency 2 to obtain the reference feedback change information 1 of the reference object 1, the reference feedback change information of all the reference objects is averaged, and finally the reference feedback change information X of the reference object is obtained.

206. The computer device determines a data label of the experimental object based on the historical feedback information, the reference feedback change information, and the feedback information.

For example, the data tags include a positive tag (which may be labeled with numeral 1) and a negative tag (which may be labeled with numeral 2), and the computer device calculates experimental feedback change information of the experimental subject based on the historical feedback information, the reference feedback change information, and the feedback information, compares the experimental feedback change information with a set threshold, determines that the data tag of the experimental subject is 1 if the experimental feedback change information is greater than the set threshold, and determines that the data tag of the experimental subject is 2 if the experimental feedback change information is not greater than the set threshold.

207. And training the prediction model by the computer equipment based on the experimental object and the data label to obtain the trained prediction model.

208. And the computer equipment determines a target object of the content push task from the full-scale objects through the trained prediction model so as to push the content through the target object.

For example, the label prediction information and the confidence level of each full-scale object are output through the trained prediction model, the full-scale object with the data label of 1 is reserved as a candidate object, ranking is carried out according to the confidence level of the label prediction information of each candidate object, then the target object is determined from the candidate objects according to the ranking information, and E-commerce advertisement is carried out on the target object.

In order to better implement the content push method provided in the embodiment of the present application, an embodiment of the present application further provides a device based on the content push method. Wherein, the meaning of the noun is the same as that in the content push method, and the specific implementation details can refer to the description in the method embodiment.

Fig. 4 is a schematic structural diagram of a content pushing apparatus provided in an embodiment of the present application, where the content pushing apparatus may include a sample obtaining module 301, an extracting module 302, a pushing module 303, a tag determining module 304, a training module 305, and an object determining module 306, where:

an obtaining module 301, configured to obtain a full amount of objects of a content push task;

An extraction module 302, configured to extract a full-scale object to obtain multiple experimental objects;

the pushing module 303 is configured to push task data to the experimental object based on the content pushing task to obtain feedback information of the experimental object on the task data, where the task data is associated with the content pushing task;

a tag determination module 304, configured to determine a data tag of the experimental object according to the feedback information;

a training module 305, configured to train the prediction model based on the experimental object and the data label, to obtain a trained prediction model;

and the object determining module 306 is configured to determine, through the trained prediction model, a target object of the content push task from the full amount of objects, so as to push the content through the target object.

In some embodiments, the feedback information includes positive feedback information and negative feedback information, the data tags include positive tags and negative tags, the tag determination module includes a first determination submodule and a second determination submodule, wherein,

the first determining submodule is used for determining the data label of the experimental object as a forward label when the feedback information of the experimental object is the forward feedback information;

and the second determining submodule is used for determining the data label of the experimental object as a negative label when the feedback information of the experimental object is negative feedback information.

In some embodiments, the content push apparatus further comprises:

the calculation unit is used for calculating the experiment feedback change information of the experimental object according to the historical feedback information, the reference feedback change information and the feedback information; (ii) a

In some embodiments, the preset threshold includes a preset first threshold and a preset second threshold, the data tag includes a positive tag, a negative tag, and a neutral tag, and the determining unit is specifically configured to:

when the experiment feedback change information of the experimental object is larger than or equal to a preset first threshold value, determining that the data label of the experimental object is a forward label;

when the experiment feedback change information of the experimental object is smaller than a preset first threshold and larger than a preset second threshold, determining that the data label of the experimental object is a neutral label;

and when the experiment feedback change information of the experiment object is less than or equal to a preset second threshold value, determining that the data label of the experiment object is a negative label.

In some embodiments, the determination submodule is specifically configured to:

when the predicted tag information of the full-scale object is a target data tag, determining the full-scale object as a candidate object;

In some embodiments, the training module is specifically configured to:

carrying out feature construction on an experimental object to obtain an experimental feature vector;

obtaining a model prediction result of an experimental object through the experimental characteristic vector and a prediction model;

In the application, the obtaining module 301 obtains a full object of a content push task, the extracting module 302 extracts the full object to obtain a plurality of experimental objects, the pushing module 303 pushes task data to the experimental objects based on the content push task to obtain feedback information of the experimental objects to the task data, the task data is associated with the content push task, the tag determining module 304 determines data tags of the experimental objects according to the feedback information, the training module 305 trains the prediction model based on the experimental objects and the data tags to obtain a trained prediction model, and finally the object determining module 306 determines a target object of the content push task from the full object through the trained prediction model to push content through the target object.

In addition, an embodiment of the present application further provides a computer device, where the computer device may be a terminal or a server, as shown in fig. 5, which shows a schematic structural diagram of the computer device according to the embodiment of the present application, and specifically:

the computer device may include components such as a processor 401 of one or more processing cores, memory 402 of one or more computer-readable storage media, a power supply 403, and an input unit 404. Those skilled in the art will appreciate that the computer device configuration illustrated in FIG. 5 does not constitute a limitation of computer devices, and may include more or fewer components than those illustrated, or some components may be combined, or a different arrangement of components. Wherein:

The processor 401 is a control center of the computer device, connects various parts of the entire computer device using various interfaces and lines, performs various functions of the computer device and processes data by operating or executing software programs and/or modules stored in the memory 402 and calling data stored in the memory 402, thereby integrally monitoring the computer device. Alternatively, processor 401 may include one or more processing cores; preferably, the processor 401 may integrate an application processor and a modem processor, wherein the application processor mainly handles operating systems, user pages, application programs, and the like, and the modem processor mainly handles wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 401.

The memory 402 may be used to store software programs and modules, and the processor 401 executes various functional applications and data processing by operating the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data created according to use of the computer device, and the like. Further, the memory 402 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 access to the memory 402.

The computer device further comprises a power supply 403 for supplying power to the various components, and preferably, the power supply 403 is logically connected to the processor 401 through a power management system, so that the functions of managing charging, discharging, and power consumption are realized through the power management system. The power supply 403 may also include any component of one or more dc or ac power sources, recharging systems, power failure detection circuitry, power converters or inverters, power status indicators, and the like.

The computer device may also include an input unit 404, the input unit 404 being operable to receive input numeric or character information and to generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

Although not shown, the computer device may further include a display unit and the like, which are not described in detail herein. Specifically, in this embodiment, the processor 401 in the computer device loads the executable file corresponding to the process of one or more application programs into the memory 402 according to the following instructions, and the processor 401 runs the application programs stored in the memory 402, thereby implementing various functions as follows:

Acquiring a full object of a content pushing task; extracting the full quantity of objects to obtain a plurality of experimental objects; pushing task data to the experimental object based on the content pushing task to obtain feedback information of the experimental object to the task data, wherein the task data is associated with the content pushing task; determining a data label of the experimental object according to the feedback information; training the prediction model based on the experimental object and the data label to obtain a trained prediction model; and determining a target object of the content push task from the full-scale objects through the trained prediction model so as to push the content through the target object.

The above operations can be implemented in the foregoing embodiments, and are not described in detail herein.

According to an aspect of the application, a computer program product or computer program is provided, comprising computer instructions, the computer instructions being stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the method provided in the various alternative implementations of the above embodiments.

It will be understood by those skilled in the art that all or part of the steps of the methods of the above embodiments may be performed by a computer program, which may be stored in a computer-readable storage medium and loaded and executed by a processor, or by related hardware controlled by the computer program.

To this end, the present application further provides a storage medium, in which a computer program is stored, where the computer program can be loaded by a processor to execute the steps in any one of the content push methods provided in the present application. For example, the computer program may perform the steps of:

acquiring a full object of a content pushing task; extracting the full-scale objects to obtain a plurality of experimental objects; pushing task data to the experimental object based on the content pushing task to obtain feedback information of the experimental object to the task data, wherein the task data is associated with the content pushing task; determining a data label of the experimental object according to the feedback information; training the prediction model based on the experimental object and the data label to obtain a trained prediction model; and determining a target object of the content push task from the full-scale objects through the trained prediction model so as to push the content through the target object.

Wherein the storage medium may include: read Only Memory (ROM), Random Access Memory (RAM), magnetic or optical disks, and the like.

Since the computer program stored in the storage medium can execute the steps in any content push method provided in the embodiment of the present application, beneficial effects that can be achieved by any content push method provided in the embodiment of the present application can be achieved, for details, see the foregoing embodiment, and are not described again here.

The content push method and device provided by the embodiment of the present application are described in detail above, and a specific example is applied in the description to explain the principle and the implementation of the present application, and the description of the embodiment is only used to help understanding the method and the core idea of the present application; meanwhile, for those skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.

Claims

1. A content pushing method, comprising:

acquiring a full object of a content pushing task;

Pushing task data to the experimental object based on the content pushing task to acquire feedback information of the experimental object on the task data, wherein the task data is associated with the content pushing task;

and determining a target object of the content push task from the full-scale objects through the trained prediction model so as to push the content through the target object.

2. The method of claim 1, wherein the feedback information comprises positive feedback information and negative feedback information, wherein the data labels comprise positive labels and negative labels, and wherein determining the data label of the subject from the feedback information comprises:

when the feedback information of the experimental object is the forward feedback information, determining that the data label of the experimental object is a forward label;

and when the feedback information of the experimental object is negative feedback information, determining that the data label of the experimental object is a negative label.

3. The method of claim 1, further comprising:

Extracting the full-scale objects to obtain a plurality of reference objects;

acquiring historical feedback information of the experimental object;

determining reference feedback change information of the reference object, wherein the reference feedback change information comprises the change degree of the feedback information of the reference object in a preset time period;

determining the data label of the experimental object according to the feedback information comprises:

4. The method of claim 3, wherein determining the data label of the subject based on the historical feedback information, the reference feedback change information, and the feedback information comprises:

calculating experimental feedback change information of the experimental object according to the historical feedback information, the reference feedback change information and the feedback information;

and determining the data label of the experimental object according to the experimental feedback change information of the experimental object.

5. The method of claim 4, wherein the data labels comprise positive labels and negative labels, and the determining the data label of the subject according to the experimental feedback change information of the subject comprises:

6. The method of claim 4, wherein the preset threshold comprises a preset first threshold and a preset second threshold, the data labels comprise a positive label and a negative label and a neutral label, and the determining the data label of the experimental subject according to the experimental feedback change information of the experimental subject comprises:

7. The method of claim 1, wherein the determining, by the trained predictive model, a target object of the content push task from the full-scale objects for content push by the target object comprises:

Performing feature construction on the full-scale object to obtain a full-scale feature vector;

inputting the full-scale feature vector into the trained prediction model to obtain label prediction information and confidence coefficient of the full-scale object;

and determining a target object of the content push task from the full-scale objects based on the label prediction information and the confidence coefficient thereof so as to push the content through the target object.

8. The method of claim 7, wherein the determining the target object of the content push task from the full-scale objects based on the tag prediction information and the confidence thereof comprises:

9. The method of any one of claims 1 to 8, wherein training a predictive model based on the subject and the data label to obtain a trained predictive model comprises:

10. A content push apparatus, comprising: