CN112541145B

CN112541145B - Page display method, device, equipment and storage medium

Info

Publication number: CN112541145B
Application number: CN201910893817.9A
Authority: CN
Inventors: 丁娇; 曾云峰; 李沛龙; 秦首科; 黄旺; 曹旭磊; 李晨瑞; 朱威; 曹金营; 杨雷; 刘琦凯; 胡均毅; 韩聪; 姜振
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2019-09-20
Filing date: 2019-09-20
Publication date: 2023-10-17
Anticipated expiration: 2039-09-20
Also published as: CN112541145A

Abstract

The application discloses a page display method, device, equipment and storage medium, and relates to the field of big data. The specific implementation scheme is as follows: acquiring a plurality of alternative display pages, wherein the display pages are associated with a first type of display conversion and at least one second type of display conversion, and the second type of display conversion probability is higher than the first type of display conversion probability; predicting the first type display conversion probability of each alternative display page according to the first type history display page matched with the first type display conversion and the second type history display page matched with the second type display conversion; and according to the prediction result, at least one target page is obtained from the plurality of alternative display pages, and the at least one target page is displayed. According to the technical scheme provided by the embodiment of the application, the accuracy of deep conversion probability estimation can be improved under the condition that the number of deep conversion positive samples is small.

Description

Page display method, device, equipment and storage medium

Technical Field

The embodiment of the application relates to a computer technology, in particular to a big data technology, and particularly relates to a page display method, device and equipment and a storage medium.

Background

With the continuous development of internet technology, various media or software developers can issue corresponding popularization information through an internet platform, for example, in the form of a presentation page, after a user looks up the popularization information, the user can finish the actions such as recommending and forwarding media articles, activating and using software, and the like, and the actions realize deep conversion of the presentation page. In fact, the final purpose of the distribution of the promotional information is also to achieve the above-mentioned deep conversion.

In order to achieve the optimal deep conversion effect, the internet platform needs to estimate the deep conversion probability of the displayed page in advance so as to display the page in a targeted manner. In the prior art, a user behavior generating deep conversion and a corresponding display page thereof are taken as positive samples, the deep conversion probability is estimated according to feature vectors extracted from the positive samples, and pages with the deep conversion probability estimated value higher than a set threshold are displayed.

The inventors found that in the process of implementing the present invention, the prior art has the following drawbacks: for popularization information providers with higher deep conversion probability, the method can achieve a good prediction effect, but for popularization information providers with lower deep conversion probability, the problem of insufficient positive sample quantity can occur, and finally, the prediction result is inaccurate.

Disclosure of Invention

The embodiment of the application provides a page display method, device, equipment and storage medium, which can improve the accuracy of deep conversion probability estimation under the condition of insufficient number of positive samples.

In a first aspect, an embodiment of the present application provides a page display method, where the method includes:

acquiring a plurality of alternative display pages, wherein the display pages are associated with a first type of display conversion and at least one second type of display conversion, and the second type of display conversion probability is higher than the first type of display conversion probability;

predicting the first type display conversion probability of each alternative display page according to the first type history display page matched with the first type display conversion and the second type history display page matched with the second type display conversion;

and according to the prediction result, at least one target page is obtained from the plurality of alternative display pages, and the at least one target page is displayed.

Optionally, predicting the first class conversion probability of each candidate presentation page according to the first class history presentation page matched with the first class presentation conversion and the second class history presentation page matched with the second class presentation conversion, including:

Acquiring a display characteristic set corresponding to each alternative display page respectively;

matching the display feature sets with the common display conversion feature set and the individual display conversion feature set respectively;

predicting the first class conversion probability of each alternative presentation page according to the matching result;

the common display conversion feature set is generated jointly according to the first type of history display pages and the second type of history display pages;

and the personality presentation conversion feature set is independently generated according to the first type of history presentation pages.

Optionally, predicting the first type of presentation conversion probability of each candidate presentation page according to the first type of history presentation page matched with the first type of presentation conversion and the second type of history presentation page matched with the second type of presentation conversion, including:

inputting each display feature set into a pre-trained first-class display conversion probability prediction model respectively, and predicting the first-class display conversion probability of each alternative display page;

the first type of presentation conversion probability prediction model is obtained through positive sample training of the first type of history presentation page and the second type of history presentation page construction.

Optionally, before acquiring the plurality of alternative presentation pages, the method further includes:

constructing a plurality of first type positive samples corresponding to the first type conversion display by using the first type history display page;

constructing a plurality of second class positive samples corresponding to each second class presentation conversion respectively by using the second class history presentation page;

generating a sample feature set corresponding to each first type positive sample and each second type positive sample;

training a multi-task learning model by using each sample feature set to obtain a first-class presentation conversion probability prediction model;

wherein the first class exhibits a transition probability prediction model comprising: the first class exhibits predictive tasks of conversion probability and the at least one second class exhibits predictive tasks of conversion probability.

Optionally, the first class exhibits a transition probability prediction model including:

the system comprises a shared hidden layer connected with an input end, a first type of presentation conversion sub-network and at least one second type of presentation conversion sub-network, wherein the first type of presentation conversion sub-network and the at least one second type of presentation conversion sub-network are respectively connected with the shared hidden layer; the output end of the first-class presentation conversion sub-network is the output end of the first-class presentation conversion probability prediction model;

The shared hidden layer is obtained by training together with the first type positive sample and a sample feature set corresponding to the second type positive sample;

the first class presentation conversion sub-network is obtained through independent training by using a sample feature set corresponding to the first class positive sample and is used for outputting a prediction result of the first class presentation conversion probability;

and the second class presentation conversion sub-network is independently trained by using a sample feature set corresponding to the second class positive sample and is used for outputting a prediction result of the second class presentation conversion probability.

Optionally, the presentation feature set or the sample feature set is a discrete value feature set;

the shared hidden layer is specifically used for: the input discrete value feature set is converted to a continuous value feature set.

Optionally, the presenting feature set or the sample feature set includes:

at least one page feature matching a presentation page, and at least one user feature corresponding to a presentation user of the presentation page.

Optionally, the first type of presentation transformation is a deep presentation transformation, and the second type of presentation transformation is a shallow presentation transformation; the deep presentation transformation achieves transformation after completion of at least one shallow presentation transformation.

Optionally, the deep layer presentation is converted into a purchase order for forming the goods presented in the presentation page, and the shallow layer presentation is converted into a purchase consultation for forming the goods presented in the presentation page;

or,

the deep presentation is converted into activating an application program presented in a presentation page, and the shallow presentation is converted into downloading the application program presented in the presentation page.

In a second aspect, an embodiment of the present application further provides a device for displaying a page, where the device includes:

the display page acquisition module is used for acquiring a plurality of alternative display pages, wherein the display pages are associated with a first type of display conversion and at least one second type of display conversion, and the second type of display conversion probability is higher than the first type of display conversion probability;

the display conversion probability prediction module is used for predicting the first type display conversion probability of each alternative display page according to the first type history display page matched with the first type display conversion and the second type history display page matched with the second type display conversion;

and the target page display module is used for acquiring at least one target page from the plurality of alternative display pages according to the prediction result and displaying the at least one target page.

Optionally, the conversion probability prediction module is presented, including:

the display feature set acquisition unit is used for acquiring display feature sets respectively corresponding to the alternative display pages;

the display feature set matching unit is used for matching each display feature set with the common display conversion feature set and the individual display conversion feature set respectively;

the conversion probability prediction unit is used for predicting the first type conversion probability of each alternative presentation page according to the matching result;

the conversion probability prediction unit is used for inputting each display characteristic set into a pre-trained first-class display conversion probability prediction model respectively, and predicting the first-class display conversion probability of each alternative display page;

Optionally, the page display device further includes:

the first type sample construction module is used for constructing a plurality of first type positive samples corresponding to the first type conversion display by using the first type history display page before acquiring a plurality of alternative display pages;

the second type sample construction module is used for constructing a plurality of second type positive samples corresponding to each second type display conversion respectively by using the second type history display page;

the sample feature set generation module is used for generating sample feature sets corresponding to the first type of positive samples and the second type of positive samples;

the prediction model acquisition module is used for training the multi-task learning model by using each sample feature set to obtain the first-class presentation conversion probability prediction model;

Optionally, the presenting feature set or the sample feature set includes:

or alternatively;

In a third aspect, an embodiment of the present application further provides an electronic device, including:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the page rendering method provided by any embodiment of the present application.

In a fourth aspect, an embodiment of the present application further provides a non-transitory computer readable storage medium storing computer instructions, where the computer instructions are configured to cause the computer to execute the page presentation method provided in any embodiment of the present application.

According to the technical scheme, the first type display conversion probability of each alternative display page is predicted according to the first type history display page matched with the first type display conversion and the second type history display page matched with the second type display conversion, and finally at least one display page is selected from the plurality of alternative display pages according to the prediction result to display, so that the prediction of the first type display conversion probability according to the first type history display page and the second type history display page is achieved, and the problem that the prediction of the first type display conversion probability is inaccurate under the condition that the first type display conversion sample is insufficient is solved.

Other effects of the above alternative will be described below in connection with specific embodiments.

Drawings

The drawings are included to provide a better understanding of the present application and are not to be construed as limiting the application. Wherein:

FIG. 1 is a flow chart of a page presentation method in a first embodiment of the application;

FIG. 2a is a flow chart of a page presentation method in a second embodiment of the present application;

FIG. 2b is a schematic diagram of a first class of model exhibiting transition probability prediction in a second embodiment of the application;

FIG. 3 is a schematic view showing the structure of a page showing apparatus in a third embodiment of the present application;

fig. 4 is a block diagram of an electronic device for implementing a page presentation method according to an embodiment of the present application.

Detailed Description

Exemplary embodiments of the present application will now be described with reference to the accompanying drawings, in which various details of the embodiments of the present application are included to facilitate understanding, and are to be considered merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

First embodiment

Fig. 1 is a flowchart of a page display method in a first embodiment of the present application, where the technical solution of the present embodiment is applicable to a case of deep conversion probability estimation according to a history display page of a first type of display conversion and a history display page of a second type of display conversion at the same time, where the method may be executed by a page display device, and the device may be implemented by software and/or hardware and may be generally integrated in a server, and the method of the present embodiment specifically includes the following steps:

Step 110, obtaining a plurality of alternative display pages, wherein the display pages are associated with a first type of display conversion and at least one second type of display conversion, and the second type of display conversion probability is higher than the first type of display conversion probability.

The conversion display probability of the first type of conversion display is the conversion display probability of the first type of conversion display, and the conversion display probability of the second type of conversion display is the conversion display probability of the second type of conversion display.

The presentation page is a page provided by a client of the Internet platform and required to be presented to a user of the Internet platform.

The internet platform may be a platform for providing a certain internet service, and illustratively, the internet platform may be a search engine, a video website or forum, a social networking site, or the like, and the client of the internet platform may be a platform having a cooperative relationship with the internet platform and pushing information through the internet platform, and illustratively, the client of the internet platform may be a media platform, a cooperative merchant (offline or online), a software developer, or the like. The user of the internet platform may be a general user who accesses the internet platform to obtain the corresponding service.

The display page can be a page which needs to be displayed in an internet platform and comprises popularization information. The promotion information can be a media article which needs to be promoted by a media platform, one or more commodities which need to be promoted by a merchant, one or more application programs which need to be promoted by software development, and the like.

In a specific application scenario of this embodiment, a user of an internet platform accesses the internet platform and satisfies a page pushing condition (for example, the access duration exceeds a set threshold, or clicks to view a certain set page, etc.), then the user may trigger to obtain multiple alternative presentation pages corresponding to a certain client of the internet platform, where different alternative presentation pages respectively carry different popularization information.

In this embodiment, first, a plurality of alternative presentation pages provided by a client of an internet platform are acquired, the presentation pages are presented on the internet platform, a user of the internet platform may click on the presentation pages and perform a series of operations, such as information consultation, registration account number, and commodity purchase, the series of operations performed by the user are called presentation conversion, and are classified into two types, namely, the presentation page is associated with the first type of presentation conversion and at least one second type of presentation conversion according to the conversion degree, wherein the first type of presentation conversion represents a presentation conversion with a higher conversion degree, such as commodity purchase, and the second type of presentation conversion represents a presentation conversion with a lower conversion degree, such as information consultation or account number registration, and in general, the internet platform user may perform the first type of presentation conversion on the basis of the second type of presentation conversion, so that the second type of presentation conversion probability is higher than the first type of presentation conversion probability.

The method includes the steps that when an alternative showing page is a commodity advertisement page provided by an online merchant, a user clicks the showing page and performs account registration to serve as second-type showing conversion, and after account registration is completed, the user performs commodity purchase, and then successful commodity ordering operation serves as first-type showing conversion.

Optionally, the first type of presentation transformation is a deep presentation transformation, and the second type of presentation transformation is a shallow presentation transformation; deep presentation transformation the transformation is achieved after completion of at least one shallow presentation transformation.

In this optional embodiment, deep presentation conversion is defined into a first type of presentation conversion, where deep presentation conversion refers to conversion with higher conversion degree, for example, an internet platform user completes a commodity ordering operation by clicking a commodity advertisement page, shallow presentation conversion refers to conversion with lower conversion degree, for example, an internet platform user performs commodity information consultation or browses a commodity webpage for more than 5 minutes by clicking a commodity advertisement page, and in general, a user performs ordering after consulting commodity information or continuously browses for a certain period of time, that is, deep presentation conversion is implemented after at least one shallow presentation conversion is completed.

or,

In the optional embodiment, specific examples of deep display conversion and shallow display conversion are provided, and when the display page is an advertisement page of the commodity, the deep display conversion is used for forming a purchase order of the commodity displayed in the display page, and the shallow display conversion is used for forming a purchase consultation of the commodity displayed in the display page; or when the display page is an application program promotion page, the deep display is converted into activating the application program displayed in the display page, and the shallow display is converted into downloading the application program displayed in the display page.

And step 120, predicting the first type display conversion probability of each alternative display page according to the first type history display page matched with the first type display conversion and the second type history display page matched with the second type display conversion.

The first type history display page matched with the first type display conversion is a page subjected to the first type display conversion, for example, a display page for forming a purchase order of the commodity displayed in the display page; the second type history presentation page matched with the second type presentation conversion is a page in which the second type presentation conversion occurs, for example, a presentation page in which purchase consultation of the goods presented in the presentation page is formed.

In this embodiment, since the client of the internet platform cooperates with the internet platform to display the number of deep display conversions, that is, to increase the number of first-class display conversions, for example, when the client of the internet platform is an online merchant, the client's final purpose is to increase the number of merchandise orders, the internet platform predicts the first display conversion probabilities of the plurality of candidate display pages in turn according to the first-class historical display pages and the second-class historical display pages, so as to determine whether to display the candidate display pages according to the first display conversion probabilities. For example, the first presentation conversion probability may be determined according to a matching result by matching feature information extracted from the alternative presentation page with feature information extracted from the first type of history presentation page and the second type of presentation page.

Optionally, predicting the first class conversion probability of each candidate presentation page according to the first class history presentation page matched with the first class presentation conversion and the second class history presentation page matched with the second class presentation conversion may be:

predicting the first type of display conversion probability of each candidate display page according to the matching result;

In this optional embodiment, a specific manner of predicting a first class conversion probability of each alternative presentation page according to a first class history presentation page and a second class history presentation page is provided, and the specific process is as follows:

firstly, a display feature set corresponding to each alternative display page is obtained, wherein the display feature set can be a feature vector, the feature vector contains the features of the current alternative display page and the features of the current user, for example, the display feature set can be (commodity category: mother and infant articles, commodity name: infant food, user age: 29, user gender: girl, user identity portrait: novice mom), and of course, the feature information in the display feature set can be replaced by an information identifier, for example, the sex is male, the sex is female, and the like.

Secondly, the display feature sets are respectively matched with a common display conversion feature set and a personalized display conversion feature set, wherein the common display conversion feature set is extracted from a first type of history display page and a second type of history display page, contains all feature information contained in the two types of history display pages, and can be extracted from one of the first type of history display pages, for example: (commodity category: outdoor supplies, commodity name: outdoor jacket, user age: 30, user gender: man, user identity image: white collar), and all feature sets extracted from all first type history display pages are formed into a feature set 1, and similarly, all feature sets extracted from all second type history display pages are formed into a feature set 2, and then the feature set 1 and the feature set 2 are jointly formed into a common display conversion feature set, wherein the personality display conversion feature set is feature information extracted from only the first type history display pages, and in the above example, all feature sets extracted from all first type history display pages are formed into a feature set 1, namely the personality display conversion feature set.

Alternatively, the feature sets in the set 1 and the set 2 may be clustered respectively to obtain multiple categories, and then the similarity between the feature set and the common display feature set and the similarity between the feature set and the individual display feature set may be calculated according to the similarity between the feature set and each category in the set 1 and the set 2, and specifically, the average value or the maximum value of the similarity between the feature set and each category in the set 1 may be obtained and used as the similarity between the feature set and the common display feature set; and acquiring the average value or the maximum value of the similarity between the display feature set and each category in the set 2, and taking the average value or the maximum value as the similarity between the display feature set and the individual display feature set.

Finally, the display feature sets corresponding to the alternative display pages are respectively matched with the common display conversion feature set and the individual display conversion feature set, the similarity between the display feature set and the common display feature set and the similarity between the display feature set and the individual display feature set can be calculated respectively, finally, the two similarity values are weighted and summed, and finally, the first type display conversion probability is obtained.

For example, the similarity between the display feature set and the common display feature set is 60%, the similarity between the display feature set and the individual display feature set is 45%, and further, the two similarities are weighted and summed to obtain the first type of display conversion probability, for example, if the weights of the common display feature set and the individual display feature set are set to be the same, the method can be according to the formula: 60% +45% + 50%, the first class calculated exhibits a probability of conversion.

Of course, it will be understood by those skilled in the art that the weights of the common display feature set and the individual display feature set may be determined according to the actual situation (for example, the number ratio of the first type of history display pages to the second type of history display pages), which is not limited in this embodiment.

And 130, acquiring at least one target page from a plurality of alternative display pages according to the prediction result, and displaying the at least one target page.

In this embodiment, according to the obtained first type of presentation conversion probability, at least one target page is obtained from a plurality of candidate presentation pages and presented, and specifically, according to a set presentation conversion probability threshold, the predicted candidate presentation page with the first type of presentation conversion probability higher than the presentation conversion probability threshold may be presented, for example, the presentation conversion probability threshold is set to be 45%; or after the alternative presentation pages with the first class presentation conversion probability higher than the presentation conversion probability threshold are acquired, the acquired alternative presentation pages are ordered according to the order of the first class presentation conversion probability from high to low, and a set number (for example, 3 or 5 or the like) of alternative presentation pages are acquired according to the order of the ordering result from high to low to be presented.

According to the technical scheme, the first type display conversion probability of each alternative display page is predicted according to the first type history display page matched with the first type display conversion and the second type history display page matched with the second type display conversion, and finally at least one display page is selected from the plurality of alternative display pages according to the prediction result to display, so that the prediction of the first type display conversion probability according to the first type history display page and the second type history display page is realized, and the problem that the prediction of the first type display conversion probability is inaccurate under the condition that the first type display conversion sample is insufficient is solved.

Second embodiment

Fig. 2a is a flowchart of a page showing method in a second embodiment of the present application, where the present embodiment is further refined on the basis of the foregoing embodiment, and specific steps of predicting a first class conversion probability of each alternative showing page and specific steps before obtaining a plurality of alternative showing pages are provided according to a first class history showing page matched with a first class showing conversion and a second class history showing page matched with a second class showing conversion. The following describes a page presentation method according to a second embodiment of the present application with reference to fig. 2a, including the following steps:

Step 210, constructing a plurality of first type positive samples corresponding to first type conversion display by using a first type history display page; and constructing a plurality of second class positive samples corresponding to each second class presentation conversion respectively by using the second class history presentation page.

The first embodiment is specifically referred to as an explanation of the first type history display page and the first type history display page, which are not described herein.

The first type of the positive sample is composed of a first type of history display page and a user generating first type of display conversion on the first type of history display page, for example, the first type of the positive sample comprises a display page forming a purchase order of goods displayed in the display page and a user corresponding to the purchase order, wherein the user can be a user ID; the second positive sample is composed of a second type history display page and users generating second type display conversion on the second type history display page, for example, the second positive sample comprises display pages forming purchase consultation of the commodities displayed in the display pages and users corresponding to the purchase consultation.

In this embodiment, a first type positive sample and a second type positive sample are respectively configured by using a first type history display page and a second type history display page, specifically, the first type positive sample is formed by the first type history display page and a user generating first type display conversion matched with the first type history display page, and the second type positive sample is formed by the second type history display page and a user generating second type display conversion matched with the second type history display page.

Step 220, generating a sample feature set corresponding to each first type positive sample and each second type positive sample.

The sample feature set is a set of feature information extracted from the first type or the second type of positive samples, and may be a feature vector with a set format.

In this embodiment, sample feature sets corresponding to each first type of positive sample and each second type of positive sample are respectively generated, and a sample feature set is formed by combining the page feature of the first type of history display page and the user feature, where the sample feature set is specifically obtained, and the finally formed sample feature set is (commodity type: outdoor article, commodity name: outdoor jacket, user age: 30, user sex: male, user identity: white collar). The generation process of the sample feature set corresponding to the second type of positive sample is the same as the generation process of the sample feature set corresponding to the first type of positive sample, and will not be further described herein.

Step 230, training the multi-task learning model by using each sample feature set to obtain a first-class presentation conversion probability prediction model;

the first class of the display conversion probability prediction model comprises the following components: the first class exhibits predictive tasks of conversion probability and the at least one second class exhibits predictive tasks of conversion probability.

The multi-task learning model is designed for the condition that training samples are insufficient, and can be used for alleviating the problem of data sparseness by using useful information of other related learning tasks. In this embodiment, the multitasking model may simultaneously estimate the first type of presentation conversion probability by using the first type of history presentation page and the second type of history presentation page under the condition that the first type of history presentation page has insufficient data.

In this embodiment, the sample feature set obtained in step 220 is input to the multi-task learning model for training to optimize parameters in the multi-task learning model, and finally, the first presentation transformation probability prediction model is obtained. The first display conversion probability prediction model comprises a prediction task of a first type display conversion probability and at least one prediction task of a second type display conversion probability, and the prediction task is used for predicting the first type display conversion probability and the second type display conversion probability respectively.

Optionally, the first class exhibits a transition probability prediction model comprising: the system comprises a shared hidden layer connected with an input end, a first type of presentation conversion sub-network and at least one second type of presentation conversion sub-network, wherein the first type of presentation conversion sub-network and the at least one second type of presentation conversion sub-network are respectively connected with the shared hidden layer; the output end of the first-class presentation conversion sub-network is the output end of the first-class presentation conversion probability prediction model;

the shared hidden layer is obtained by training a sample feature set corresponding to the first type of positive sample and the second type of positive sample together;

the first class presentation conversion sub-network is obtained by independent training by using a sample feature set corresponding to the first class positive sample and is used for outputting a prediction result of the first class presentation conversion probability;

and the second class presentation conversion sub-network is obtained by independent training by using a sample feature set corresponding to the second class positive sample and is used for outputting a prediction result of the second class presentation conversion probability.

In this optional embodiment, a specific structure of a first type of presentation transformation probability prediction model is provided, as shown in fig. 2b, including a shared hidden layer connected to an input end, a first type of presentation transformation sub-network respectively connected to the shared hidden layer, and at least one second type of presentation transformation sub-network; the output end of the first-class presentation conversion sub-network is the output end of the first-class presentation conversion probability prediction model.

The shared hidden layer is obtained by training sample feature sets corresponding to the first type positive samples and the second type positive samples together, namely, parameters in the shared hidden layer are optimized through the sample feature sets corresponding to the first type positive samples and the second type positive samples. The first type of presentation conversion sub-network is obtained through first type of positive sample training, and can predict first type of presentation conversion probability according to the input presentation characteristic set. The second-class presentation conversion sub-network is obtained through second-class positive sample training, and can predict second-class presentation conversion probability according to the input presentation characteristic set.

Step 240, obtaining a plurality of alternative display pages, wherein the display pages are associated with a first type of display conversion and at least one second type of display conversion, and the display conversion probability of the second type of display conversion is higher than that of the first type of display conversion.

Step 250, obtaining display feature sets respectively corresponding to the candidate display pages.

In this embodiment, a display feature set corresponding to each candidate display page is obtained, where the display feature set may be a feature vector, and the feature vector includes features of the current candidate display page and features of the current user.

Optionally, the display feature set or the sample feature set is a discrete value feature set;

In this optional embodiment, the above-mentioned display feature set is a set formed by feature information of each display page, and the sample feature set is a set formed by feature information of each first type of history display page and each second type of history display page, so that the sample feature set is a discrete value feature set. Illustratively, discretizing the age information of the user in the plurality of sample feature sets yields discrete value features of the ages, for example, (0 years, 10 years) set to 1, (10 years, 20 years) set to 2, (20 years, 30 years) set to 3, and so on. Through the discretization process described above, much information is actually lost because all age characteristics (20 years, 30 years) use 2 as a discretization result, in order to compensate for the discretization loss described above in the model, because the input discrete value characteristic set can be converted into a continuous value characteristic set at the shared hidden layer.

Optionally, presenting the feature set or the sample feature set includes:

at least one page feature matching the presentation page, and at least one user feature corresponding to a presentation user of the presentation page.

In this optional embodiment, specific feature information contained in the display feature set and the sample feature set is provided, including at least one page feature matched with the display page and at least one user feature corresponding to a display user of the display page, where, by way of example, when the display page is an application promotion page, the page feature matched with the display page includes (application category: traffic, application name: map), the user feature corresponding to the display user of the display page includes (user age: 32, user gender: male, user identity: driver), and the two features together form the display feature set (application category: traffic, application name: map, user age: 32, user gender: male, user identity: driver).

260, respectively inputting each display feature set into a pre-trained first-class display conversion probability prediction model, and predicting the first-class display conversion probability of each alternative display page;

the first type of presentation conversion probability prediction model is obtained by training positive samples constructed by the first type of history presentation pages and the second type of history presentation pages.

In this embodiment, the acquired display feature set of each candidate display page is input into a pre-trained first-type display conversion probability prediction model, the first-type display conversion probability of the current candidate display page is predicted, and finally the output information of the first-type display conversion probability prediction model is the first-type display conversion probability of the display page.

Step 270, according to the prediction result, at least one target page is obtained from the multiple candidate display pages, and the at least one target page is displayed.

According to the technical scheme, a first type positive sample and a second type positive sample are constructed according to a first type history display page and a second type history display page, then a sample feature set is generated according to the two types of positive samples, a multi-task learning model is trained by using the generated sample feature set to obtain a first type display conversion probability prediction model, the first type display conversion probability is predicted by inputting a display feature set corresponding to an alternative display page into the first type display conversion probability prediction model, at least one display page is selected from the plurality of alternative display pages according to a prediction result to display, the prediction of the first type display conversion probability by adopting the pre-trained first type display conversion probability prediction model is realized, and the prediction quality of the first type display conversion probability is improved.

Third embodiment

Fig. 3 is a schematic structural diagram of a page display device according to a third embodiment of the present application, where the page display device includes: a presentation page acquisition module 310, a presentation conversion probability prediction module 320, and a target page display module 330.

A presentation page obtaining module 310, configured to obtain a plurality of alternative presentation pages, where the presentation pages are associated with a first type of presentation conversion and at least one second type of presentation conversion, and a second type of presentation conversion probability is higher than the first type of presentation conversion probability;

the display conversion probability prediction module 320 is configured to predict a first type display conversion probability of each of the candidate display pages according to a first type history display page that matches the first type display conversion and a second type history display page that matches the second type display conversion;

and the target page display module 330 is configured to obtain at least one target page from the plurality of candidate presentation pages according to the prediction result, and present the at least one target page.

Optionally, the conversion probability prediction module 320 is presented, including:

Optionally, the page display device further includes:

Optionally, the presenting feature set or the sample feature set includes:

or alternatively;

The page display device provided by the embodiment of the application can execute the page display method provided by any embodiment of the application, and has the corresponding functional modules and beneficial effects of the execution method.

Fourth embodiment

According to an embodiment of the present application, the present application also provides an electronic device and a readable storage medium.

Fig. 4 is a block diagram of an electronic device according to a page presentation method according to an embodiment of the present application. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the applications described and/or claimed herein.

As shown in fig. 4, the electronic device includes: one or more processors 401, memory 402, and interfaces for connecting the components, including a high-speed interface and a low-speed interface. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions executing within the electronic device, including instructions stored in or on memory to display graphical information of the GUI on an external input/output device, such as a display device coupled to the interface. In other embodiments, multiple processors and/or multiple buses may be used, if desired, along with multiple memories and multiple memories. Also, multiple electronic devices may be connected, each providing a portion of the necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). One processor 401 is illustrated in fig. 4.

Memory 402 is a non-transitory computer readable storage medium provided by the present application. The memory stores instructions executable by at least one processor to cause the at least one processor to perform the page presentation method provided by the present application. The non-transitory computer readable storage medium of the present application stores computer instructions for causing a computer to execute the page presentation method provided by the present application.

The memory 402 is used as a non-transitory computer readable storage medium for storing non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the page presentation method in the embodiment of the present application (e.g., the presentation page acquisition module 310, the presentation conversion probability prediction module 320, and the target page display module 330 shown in fig. 3). The processor 401 executes various functional applications of the server and data processing, i.e., implements the page presentation method in the above-described method embodiments, by running non-transitory software programs, instructions, and modules stored in the memory 402.

Memory 402 may include a storage program area that may store an operating system, at least one application program required for functionality, and a storage data area; the storage data area may store data created from the use of the page rendering electronic device, and the like. In addition, memory 402 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 402 may optionally include memory located remotely from processor 401, which may be connected to the page rendering electronic device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The electronic device of the page presentation method may further include: an input device 403 and an output device 404. The processor 401, memory 402, input device 403, and output device 404 may be connected by a bus or otherwise, for example in fig. 4.

The input device 403 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the page rendering electronic device, such as a touch screen, keypad, mouse, trackpad, touchpad, pointer stick, one or more mouse buttons, trackball, joystick, and like input devices. The output device 404 may include a display apparatus, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors), among others. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device may be a touch screen.

Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASIC (application specific integrated circuit), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

These computing programs (also referred to as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented in a high-level procedural and/or object-oriented programming language, and/or in assembly/machine language. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

According to the technical scheme provided by the embodiment of the application, the first type display conversion probability of each alternative display page is predicted by acquiring a plurality of alternative display pages and according to the first type history display page matched with the first type display conversion and the second type history display page matched with the second type display conversion, and finally at least one display page is selected from the plurality of alternative display pages according to the prediction result, so that the prediction of the first type display conversion probability according to the first type history display page and the second type history display page is realized, and the problem that the prediction of the first type display conversion probability is inaccurate under the condition that the first type display conversion sample is insufficient is solved.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present application may be performed in parallel, sequentially, or in a different order, provided that the desired results of the disclosed embodiments are achieved, and are not limited herein.

The above embodiments do not limit the scope of the present application. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the scope of the present application.

Claims

1. The page display method is characterized by comprising the following steps of:

acquiring a plurality of alternative display pages, wherein the display pages are associated with a first type of display conversion and at least one second type of display conversion, and the probability of the second type of display conversion is higher than that of the first type of display conversion, wherein the first type of display conversion is a deep display conversion, and the second type of display conversion is a shallow display conversion; the deep layer presentation conversion realizes conversion after at least one shallow layer presentation conversion is completed;

according to the prediction result, at least one target page is obtained from the plurality of alternative display pages, and the at least one target page is displayed;

the predicting the first type of display conversion probability of each alternative display page according to the first type of history display page matched with the first type of display conversion and the second type of history display page matched with the second type of display conversion comprises the following steps:

The personality presentation conversion feature set is independently generated according to the first type of history presentation pages;

wherein, the matching process is:

respectively calculating the similarity of the display feature set and the common display feature set and the similarity of the display feature set and the individual display feature set;

and carrying out weighted summation on the two similarity values to finally obtain the first-class display conversion probability.

2. The method of claim 1, wherein the deep presentation is converted to form a purchase order for the merchandise presented in the presentation page, and the shallow presentation is converted to form a purchase consultation for the merchandise presented in the presentation page;

or,

3. The page display method is characterized by comprising the following steps of:

the first type of display conversion probability prediction model is obtained by training positive samples constructed by the first type of history display pages and the second type of history display pages;

Before acquiring the plurality of alternative presentation pages, the method further comprises the following steps:

constructing a plurality of first type positive samples corresponding to the first type display conversion by using the first type history display page;

wherein the first class exhibits a transition probability prediction model comprising: a first class of predictive tasks exhibiting transition probabilities, and at least one second class of predictive tasks exhibiting transition probabilities;

wherein the first class exhibits a transition probability prediction model comprising:

4. A method according to claim 3, wherein the presentation feature set or the sample feature set is a discrete value feature set;

5. A method according to claim 3, wherein the presentation feature set or the sample feature set comprises:

6. The method of claim 3, wherein the deep presentation is converted to form a purchase order for the merchandise presented in the presentation page and the shallow presentation is converted to form a purchase consultation for the merchandise presented in the presentation page;

Or,

7. A page presentation apparatus, comprising:

the display page acquisition module is used for acquiring a plurality of alternative display pages, wherein the display pages are associated with a first type of display conversion and at least one second type of display conversion, and the probability of the second type of display conversion is higher than that of the first type of display conversion, wherein the first type of display conversion is a deep display conversion, and the second type of display conversion is a shallow display conversion; the deep layer presentation conversion realizes conversion after at least one shallow layer presentation conversion is completed;

the target page display module is used for acquiring at least one target page from the plurality of candidate display pages according to the prediction result and displaying the at least one target page;

Wherein, reveal the conversion probability prediction module, include:

wherein, the matching process is:

8. A page presentation apparatus, comprising:

wherein, reveal the conversion probability prediction module, include:

wherein, the page presentation device further includes:

The first type sample construction module is used for constructing a plurality of first type positive samples corresponding to the first type display conversion by using the first type history display page before acquiring a plurality of alternative display pages;

9. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

10. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-6.