CN112288146A

CN112288146A - Page display method, device, system, computer equipment and storage medium

Info

Publication number: CN112288146A
Application number: CN202011103373.3A
Authority: CN
Inventors: 杜永志; 徐夙龙; 彭长平; 朱鑫; 魏圣磊; 毛艺; 赵义; 兰汐; 丛冠男; 杨皑; 许佳; 王雪
Original assignee: Beijing Wodong Tianjun Information Technology Co Ltd
Current assignee: Beijing Wodong Tianjun Information Technology Co Ltd
Priority date: 2020-10-15
Filing date: 2020-10-15
Publication date: 2021-01-29

Abstract

The embodiment of the application provides a page display method, a device, a system, a model training method, computer equipment and a storage medium, which relate to the field of Internet and deep learning and comprise the following steps: generating bidding information of each target article according to the conversion cost and the conversion rate, and sending the bidding information of each target article to the electronic device, wherein the bidding information of each target article is used for sequencing the target articles to obtain a sequencing result, the sequencing result is used for outputting a page comprising the target articles, the bidding information is generated from two dimensions of the conversion cost and the conversion rate, the sequencing result is determined based on the bidding information, and the page is output, so that the bidding information is highly attached to the target articles, the sequencing result meets the requirement of a user for conveniently and quickly selecting the target articles, the click rate of the user can be increased by the output page, and the technical effect of increasing the utilization rate of network resources is achieved.

Description

Page display method, device, system, computer equipment and storage medium

Technical Field

The embodiment of the application relates to the field of internet and deep learning, in particular to a page display method, a device, a system, a model training method, computer equipment and a storage medium.

Background

With the development of the internet, the shopping convenience of people is greatly improved by online shopping, for example, a user can provide a page including each item based on an online shopping platform to select the corresponding item.

In the prior art, a merchant having an item can bid on the traffic of a network platform, the network platform generates and outputs a page including each item based on bidding information corresponding to the bidding, and the bidding information is generally obtained by means of history, online investigation, experiment and the like.

However, how to facilitate users to conveniently and quickly obtain the articles meeting the needs of the users becomes a problem to be solved urgently, and the click rate of the articles is improved.

Disclosure of Invention

The embodiment of the application provides a page display method, a page display device, a page display system, a model training method, computer equipment and a storage medium, which are used for solving the problem that a user can conveniently and quickly obtain articles meeting the requirements of the user.

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

generating bidding information of each target item according to the conversion cost and the conversion rate;

sending the bidding information of each target item to an electronic device, wherein the bidding information of each target item is used for sequencing the target items to obtain a sequencing result, and the sequencing result is used for outputting a page comprising the target items.

In this embodiment, the bid information is generated from two dimensions of the conversion cost and the conversion rate, the ranking result is determined based on the bid information, and the page is output, so that the bid information is highly attached to the target object, the ranking result meets the requirement of the user for conveniently and quickly selecting the target object, the click rate of the user can be increased through the output page, and the technical effect of increasing the utilization rate of network resources is achieved.

In some embodiments, the conversion cost is derived based on preset base cost information and cost float information.

In some embodiments, the method further comprises:

acquiring first attribute information of the target object, inputting the first attribute information into the conversion cost model for prediction, and generating the cost floating information; alternatively, the first and second electrodes may be,

and acquiring attribute information of at least one other article, inputting the attribute information of the other article into the conversion cost model for prediction, generating a mean value, and taking the mean value as the cost floating information.

In this embodiment, the cost floating information may be generated based on the first attribute information, or the average value may be determined as the cost floating information, thereby achieving flexibility and diversity in determining the cost floating information.

Preferably, if the target object has corresponding first attribute information, the cost floating information is determined based on the first attribute information, and if the target object is an object which is on-line for the first time and has no first attribute information, the average value is determined as the cost floating information, so that the technical effects of reliability and accuracy of the cost floating information are achieved.

In some embodiments, the first attribute information includes brand information and/or category information; inputting the first attribute information into a preset conversion cost model for prediction to generate the cost floating information;

acquiring a brand feature vector of the brand information, and/or acquiring a category feature vector of the category information;

and predicting according to the brand feature vector and/or the category feature vector to generate the cost floating information.

In some embodiments, the method further comprises:

acquiring second attribute information of the target object;

and inputting the second attribute information into a conversion rate model for prediction to generate the conversion rate.

In some embodiments, inputting the second attribute information to a conversion rate model for prediction, generating the conversion rate, comprises:

performing scattering operation on the second attribute information according to at least one preset time sequence;

and predicting the scattered second attribute information to generate the conversion rate.

In this embodiment, the second attribute information is broken up in a time series manner, so that the breaking efficiency can be improved, the conversion rate generation efficiency can be realized, and the technical effect of the page output efficiency can be realized.

In some embodiments, predicting the second property information after scattering to generate the conversion rate comprises:

performing remainder calculation on the scattered second attribute information to obtain a predicted data path;

and predicting the predicted data path to generate the conversion rate.

In some embodiments, generating bid information for each target item based on conversion cost and conversion rate includes:

generating an adjusting parameter according to the consumption information of the target object in a preset time period;

and generating the bidding information according to the conversion cost, the conversion rate and the adjusting parameter.

In this embodiment, the bidding information is generated by combining the adjustment parameter, the conversion cost and the conversion rate, so that the accuracy and the reliability of the bidding information can be realized, the reliability of the output page can be further realized, and the technical effects of improving the click rate and the resource utilization rate can be achieved.

In some embodiments, the consumption information includes predicted consumption information and actual consumption information.

In some embodiments, the conversion cost model is obtained by inputting the attribute information of the article to be trained into a feed-forward neural network model for training.

In a second aspect, the present embodiment further provides a page display method, where the method includes:

obtaining bidding information of each target item to be displayed, wherein the bidding information is generated based on conversion cost and conversion rate;

sorting the target items according to the bidding information;

outputting a page including the target item based on the sorting result.

In some embodiments, the conversion cost is obtained based on preset base cost information and cost float information.

In some embodiments, the cost float information is generated by inputting the first attribute information of the target item into a conversion cost model for prediction; or the cost floating information is a preset average value, wherein the preset average value is generated by inputting the attribute information of at least one other article to the conversion cost model for prediction.

In some embodiments, if the first attribute information includes brand information and/or category information, the cost float information is generated by the conversion cost model predicting a brand feature vector and/or a category feature vector;

wherein the brand feature vector is generated by the conversion cost model based on the brand information, and the category feature vector is generated by the conversion cost model based on the category information.

In some embodiments, the conversion rate is generated by inputting the second attribute information of the target item into a preset conversion rate model for prediction.

In some embodiments, the conversion rate is obtained by predicting second attribute information after scattering based on the conversion rate model, and the second attribute information after scattering is obtained by scattering the second attribute information in at least one preset time sequence based on the conversion rate model.

In some embodiments, the conversion rate is predicted based on the conversion rate model, and the predicted data path is obtained by performing a complementary calculation on the scattered second attribute information.

In some embodiments, the bid information is generated based on the conversion cost, the conversion rate, and preset tuning parameters;

wherein the adjustment parameter is generated based on consumption information of the target item within a preset time period.

In a third aspect, this embodiment provides a model training method applied to page display, where the method includes:

acquiring attribute information of a sample article;

scattering the attribute information of the sample object according to a preset deep cross network model and at least one preset training time sequence to obtain the attribute information of the scattered sample object;

training the attribute information of the scattered sample articles according to the deep cross network model to obtain a conversion rate model; the conversion rate model is used for generating a conversion rate, the conversion rate and the conversion cost are used for generating bidding information of each target item, and the bidding information is used for obtaining a sequencing result of the target items.

In some embodiments, training the attribute information of the scattered sample articles according to the deep cross network model to obtain a conversion rate model, includes:

determining a training data path according to the attribute information of the scattered sample articles;

and training according to the training data path, and generating and outputting the conversion cost model.

In some embodiments, determining a training data path from the attribute information of the broken sample item includes: and obtaining the training data path from the sample attribute information of the scattered sample articles based on a remainder mode.

determining a sample characteristic vector of the attribute information of the scattered sample article;

generating a sample prediction result corresponding to the sample feature vector;

and adjusting the deep cross network model according to the sample prediction result to generate the conversion rate model.

In some embodiments, adjusting the deep cross network model according to the sample prediction result to generate the conversion rate model comprises:

generating a sample loss function of the sample prediction result and a preset sample calibration result;

and adjusting the cross network model according to the sample loss function, and generating and outputting the conversion cost model.

In a fourth aspect, the present embodiment provides a page display apparatus, including:

the generating module is used for generating bidding information of each target item according to the conversion cost and the conversion rate;

and the sending module is used for sending the bidding information of each target item to the electronic equipment, wherein the bidding information of each target item is used for sequencing the target items to obtain a sequencing result, and the sequencing result is used for outputting a page comprising the target items.

In some embodiments, the apparatus further comprises:

the first obtaining module is used for obtaining first attribute information of the target object, inputting the first attribute information into the conversion cost model for prediction, and generating the cost floating information; alternatively, the first and second electrodes may be,

the first obtaining module is configured to obtain attribute information of at least one other article, input the attribute information of the other article to the conversion cost model for prediction, generate a mean value, and use the mean value as the cost floating information.

In some embodiments, the first obtaining module is configured to obtain a brand feature vector of the brand information and/or obtain a category feature vector of the category information; and predicting according to the brand feature vector and/or the category feature vector to generate the cost floating information.

In some embodiments, the first obtaining module is configured to obtain second attribute information of the target item, and input the second attribute information to a conversion rate model for prediction, so as to generate the conversion rate.

In some embodiments, the first obtaining module is configured to break up the second attribute information according to at least one preset time sequence, predict the broken up second attribute information, and generate the conversion rate.

In some embodiments, the first obtaining module is configured to perform remainder calculation on the scattered second attribute information to obtain a predicted data path, predict the predicted data path, and generate the conversion rate.

In some embodiments, the generating module is configured to generate an adjustment parameter according to consumption information of the target item within a preset time period, and generate the bidding information according to the conversion cost, the conversion rate, and the adjustment parameter.

In a fifth aspect, the present embodiment provides a page display apparatus, including:

the second obtaining module is used for obtaining bidding information of each target object to be displayed, wherein the bidding information is generated based on the conversion cost and the conversion rate;

the ordering module is used for ordering the target articles according to the bidding information;

and the output module is used for outputting the page comprising the target item based on the sequencing result.

In a sixth aspect, this embodiment provides a model training apparatus applied to page display, including:

the third acquisition module is used for acquiring the attribute information of the sample article;

the scattering module is used for scattering the attribute information of the sample object according to a preset deep cross network model and at least one preset training time sequence to obtain the attribute information of the scattered sample object;

the training module is used for training the attribute information of the scattered sample articles according to the deep cross network model to obtain a conversion rate model; the conversion rate model is used for generating a conversion rate, the conversion rate and the conversion cost are used for generating bidding information of each target item, and the bidding information is used for obtaining a sequencing result of the target items.

In some embodiments, the scattering module is configured to determine a training data path according to attribute information of the scattered sample article, perform training according to the training data path, and generate and output the conversion cost model.

In some embodiments, the breaking module is configured to obtain the training data path from the sample attribute information of the broken sample item based on a remainder approach.

In some embodiments, the training module is configured to determine a sample feature vector of the attribute information of the scattered sample article, generate a sample prediction result corresponding to the sample feature vector, and adjust the deep cross network model according to the sample prediction result to generate the conversion rate model.

In some embodiments, the training module is configured to generate a sample loss function of the sample prediction result and a preset sample calibration result, adjust the cross network model according to the sample loss function, and generate and output the conversion cost model.

In a seventh aspect, this embodiment provides a page display system, including:

the apparatus as described in the fourth embodiment;

the apparatus as described in the fifth embodiment.

In some embodiments, the system further comprises: the apparatus as in the sixth embodiment.

In an eighth aspect, the present embodiment provides a computer device, including: a memory, a processor;

a memory; a memory for storing the processor-executable instructions;

wherein the processor is configured to perform the method of the first embodiment; alternatively, the first and second electrodes may be,

the processor is configured to perform the method of the second embodiment; alternatively, the first and second electrodes may be,

the processor is configured to perform the method according to the third embodiment.

In a ninth aspect, the present embodiment provides a computer-readable storage medium having stored thereon computer-executable instructions for implementing the method according to the first embodiment when executed by a processor; alternatively, the first and second electrodes may be,

the computer executable instructions when executed by a processor are for implementing the method as described in the second embodiment; alternatively, the first and second electrodes may be,

the computer executable instructions, when executed by a processor, are for implementing the method as described in the third embodiment.

The embodiment of the application provides a page display method, a page display device, a page display system, a model training method, computer equipment and a storage medium, wherein the page display method comprises the following steps: according to the conversion cost and the conversion rate, generating bidding information of each target item, and sending the bidding information of each target item to the electronic device, wherein the bidding information of each target item is used for sequencing the target items to obtain a sequencing result, and the sequencing result is used for outputting a page including the target items. The method has the advantages that the problems of low putting effect and low click rate of the target object are solved, so that the determined bidding information, user feedback and high fit among the target objects are improved, the reliability and accuracy of the bidding information representing the target object are improved, the output is realized, the user can quickly and conveniently select the page of the target object meeting the requirement of the user, the click rate is improved, and the utilization rate of network resources (such as flow resources) is improved.

Drawings

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

Fig. 1 is a schematic view of an application scenario of a page display method according to an embodiment of the present application;

FIG. 2 is a first embodiment according to the present application;

FIG. 3 is a second embodiment according to the present application;

FIG. 4 is a third embodiment according to the present application;

FIG. 5 is a fourth embodiment according to the present application;

FIG. 6 is a fifth embodiment according to the present application;

FIG. 7 is a sixth embodiment according to the present application;

FIG. 8 is a seventh embodiment according to the present application;

FIG. 9 is an eighth embodiment according to the present application;

FIG. 10 is a ninth embodiment according to the present application;

FIG. 11 is a tenth embodiment according to the present application;

with the foregoing drawings in mind, certain embodiments of the disclosure have been shown and described in more detail below. These drawings and written description are not intended to limit the scope of the disclosed concepts in any way, but rather to illustrate the concepts of the disclosure to those skilled in the art by reference to specific embodiments.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

The terms referred to in the embodiments of the present application are explained as follows:

cost of transformation (Cost per Action, CPA): refers to the cost per action, i.e., the cost of action taken on the web advertisement by each visitor to the item. For example, the cost paid by the advertiser for each action.

Conversion (Conversion Rate, CVR): is an index for measuring the advertising effectiveness of an article, and can be understood as the conversion rate from a user clicking on an advertisement to a user who is effectively activated or registered or even paid for.

Article: it refers to various things or sporadic articles, such as commodities (including physical commodities and virtual commodities), and may also include material data.

Bidding information: refers to information relating to the value of an item determined in a competitive manner.

An electronic device: the device is composed of electronic components such as an integrated circuit, a transistor, an electron tube and the like, and is a device which plays a role by applying electronic technology (including) software, and comprises an electronic computer, a robot controlled by the electronic computer, a numerical control or program control system and the like.

Referring to fig. 1, fig. 1 is a schematic view illustrating an application scenario of a page display method according to an embodiment of the present application.

As shown in fig. 1, the application scenario includes: a terminal device providing services (advertisement services, push services, etc.), a plurality of merchants (such as merchant 1, merchant 2, and merchant n shown in fig. 1, and the plurality of merchants may be a plurality of advertisers).

Illustratively, the terminal device may include a server and a display, and the server may be configured to receive bid information sent by each advertiser based on the respective server, rank the advertisement information provided by each advertiser based on the bid information, obtain a ranking result, generate a page including each advertisement information according to the ranking result, and control the display to display the page (e.g., advertisement information a, advertisement information B, and advertisement information C shown in fig. 1).

Of course, in some examples, the terminal device may also determine bid information corresponding to each advertisement information, rank the advertisement information provided by each advertiser based on the bid information, obtain a ranking result, generate a page including each advertisement information according to the ranking result, and control the display to display the page.

In the related art, for example, a terminal device generates bidding information, and the terminal device generally determines the bidding information by using a history, an online survey, a test, and the like. Taking the example that the terminal device determines the bidding information based on the adoption history record:

and determining advertisement putting effects, such as click rate, evaluation and the like, corresponding to the advertisement information according to the history, and determining corresponding bidding information according to the click rate and the average person, wherein if the advertisement information A is determined to have a putting effect relatively higher than that of the advertisement information B according to the click rate, the bidding information of the advertisement information A is determined to be relatively higher than that of the advertisement information B.

However, the bid information may not accurately express the advertisement delivery effect of the advertisement information in an artificial manner, so that the user cannot conveniently and quickly obtain the advertisement information meeting the demand of the user, the click rate of the advertisement information is low, and the utilization rate of network resources is relatively low.

The inventor of the application obtains the inventive concept of the application through creative work: the bid information is automatically determined from two dimensions of the conversion cost and the conversion rate, and the corresponding page is output based on the bid information, so that the user can conveniently and quickly determine the articles meeting the needs of the user, the click rate is improved, and the utilization rate of network resources is improved.

The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.

Fig. 2 is a diagram illustrating a page display method according to a first embodiment of the present application, as shown in fig. 2, the page display method includes:

s101: and generating bidding information of each target item according to the conversion cost and the conversion rate.

For example, the execution main body of the embodiment may be a page display device, and the page display device may be a server (including a local server and a cloud server), a terminal device, a processor, a chip, and the like, which is not limited in the embodiment.

If the page display method of the present embodiment is applied to the application scenario shown in fig. 1, the page display apparatus may be the terminal device shown in fig. 1.

In this embodiment, taking the target object as the advertisement information to be displayed (which may be advertisement information for a certain product) as an example, the step may be understood as: one advertisement information to be displayed corresponds to one conversion cost and one conversion rate, and the bidding information of the advertisement information to be displayed can be determined according to the conversion cost and the conversion rate of the certain advertisement information to be displayed.

It is worth to be noted that, in this embodiment, a scheme for determining bid information from two dimensions is introduced, and specifically, bid information is determined through two dimensions of a conversion cost and a conversion rate, so as to avoid the problems that, due to the fact that bid information is determined through tests, manual work and other manners in the related art, the bid information cannot accurately represent relevant features of advertisement information (such as features of user clicks and features of user evaluations) due to subjective factors, resulting in a low advertisement information delivery effect and a low click rate, and by determining the bid information from two dimensions in this embodiment, high adhesion among the determined bid information, advertisement information and user feedback (such as clicks and evaluations) can be improved, thereby improving reliability and accuracy of representing advertisement information by the bid information, and further realizing that when a page including advertisement information is output based on the bid information, the user can quickly and conveniently select the advertisement information meeting the demand of the user, the click rate is improved, and the utilization rate of network resources (such as flow resources) is improved.

S102: and sending the bidding information of each target item to the electronic equipment, wherein the bidding information of each target item is used for sequencing the target items to obtain a sequencing result, and the sequencing result is used for outputting a page comprising the target items.

Based on the above example, after determining the bid information of each advertisement information to be displayed, the page display device may send the bid information of each advertisement information to be displayed to the electronic device, and the electronic device may rank the advertisements to be displayed based on the bid information of the advertisements to be displayed, and output a page including the advertisements to be displayed according to a ranking result obtained by ranking.

Illustratively, the sorting may be in ascending order or in descending order. The following is exemplarily described in ascending order:

the electronic device may rank the advertisement information to be displayed in a descending order based on the bid information to obtain a ranking result, and in the output page, the last advertisement information to be displayed in the ranking result is displayed in front of other advertisement information, and so on.

Similarly, the following is exemplarily described in ascending order of the sort:

the electronic equipment can sort the advertisement information to be displayed in a descending order based on the bidding information to obtain a sorting result, and in the output page, the advertisement information to be displayed which is sorted at the top in the sorting result is displayed in front of other advertisement information, and so on.

Based on the above analysis, the present embodiment provides a page display method, including: according to the conversion cost and the conversion rate, generating bidding information of each target item, and sending the bidding information of each target item to the electronic device, wherein the bidding information of each target item is used for sequencing the target items to obtain a sequencing result, and the sequencing result is used for outputting a page including the target items. The method has the advantages that the problems of low putting effect and low click rate of the target object are solved, so that the determined bidding information, user feedback and high fit among the target objects are improved, the reliability and accuracy of the bidding information representing the target object are improved, the output is realized, the user can quickly and conveniently select the page of the target object meeting the requirement of the user, the click rate is improved, and the utilization rate of network resources (such as flow resources) is improved.

Fig. 3 is a diagram illustrating a page display method according to a second embodiment of the present application, as shown in fig. 3, the page display method includes:

s201: and acquiring first attribute information of the target object, inputting the first attribute information into a preset conversion cost model for prediction, and generating cost floating information.

For example, the first attribute information may be used to characterize information related to the target item, such as picture information of the target item, text information of the target item, and so on.

For example, the cost floating information may be used to represent information of the cost floating up and down, such as information of the floating interval of the cost.

Based on the above mentioned target object being the advertisement information to be displayed as an example, the step may be understood as obtaining the content of the advertisement information to be displayed, which may include a picture of the advertisement or may include a text of the advertisement.

For example, the picture of the advertisement information to be displayed may be a picture of the advertisement itself (for example, if the advertisement is an advertisement for a mobile phone, the picture of the advertisement itself may be understood as a picture of the mobile phone), the picture of the advertisement may also be a picture of a user feedback for the advertisement (for example, a real object of the mobile phone taken by the user), and so on, which are not listed here any more. The text of the advertisement may be a description text of the advertisement, a comment text of the advertisement, etc., and is not listed here.

In some embodiments, a method of constructing a conversion cost model may include:

step 1: and acquiring an article to be trained.

Illustratively, the items to be trained are used for characterization for training the items that generate the conversion cost model.

It should be noted that the selection of the type and the number of the to-be-trained objects in the present embodiment is not limited, and the selection may be set by the page display device based on the requirement, history, experiment, and the like.

Step 2: and acquiring attribute information of the object to be trained.

Similarly, the attribute information of the to-be-trained object may be used for representing, and describing related information of the to-be-trained object, such as picture information and text information.

And step 3: and inputting the attribute information of the object to be trained into the feedforward neural network model to generate a conversion cost model.

In some embodiments, step 3 may comprise: generating training characteristic vectors of attribute information of an object to be trained based on a feedforward neural network model, training the training characteristic vectors by combining parameters (such as weight coefficients, channel quantity and the like) of the feedforward neural network model, generating a training result (cost floating information obtained by training), comparing the training result with a preset training calibration result (cost floating information calibrated in advance), generating a training loss function of the training result and the training calibration result, adjusting the feedforward neural network model according to the training loss function, specifically adjusting the parameters of the feedforward neural network model until the training loss function meets a preset training threshold or the training iteration number reaches a preset number threshold, and determining the corresponding feedforward neural network model as a conversion cost model.

Illustratively, the first attribute information includes brand information and/or category information, the brand information may be used for representing information related to a manufacturer of the target object, and the category information may be used for representing information related to a category to which the target object belongs (such as household appliances, kitchen ware and the like).

If the first attribute information comprises brand information, inputting the first attribute information into a conversion cost model for prediction, and generating cost floating information, wherein the cost floating information comprises: and determining a feature vector of the brand information based on the conversion cost model, predicting the feature vector of the brand information based on the conversion cost model, and generating cost floating information.

If the first attribute information comprises category information, inputting the first attribute information into a conversion cost model for prediction, and generating cost floating information, wherein the method comprises the following steps: determining the feature vector of the category information based on the conversion cost model, predicting the feature vector of the category information based on the conversion cost model, and generating cost floating information.

If the first attribute information comprises brand information and category information, inputting the first attribute information into a conversion cost model for prediction, and generating cost floating information, wherein the cost floating information comprises: determining a feature vector of the brand information based on the conversion cost model, determining a feature vector of the category information based on the conversion cost model, predicting the feature vector of the brand information and the feature vector of the category information based on the conversion cost model, and generating cost floating information.

Illustratively, in some embodiments, the first attribute information may further include an item characteristic of the target item, which may include: the price of the target item, the comment on the target item (including at least one of the quantity, the content, the goodness proportion, etc.), the distribution mode of the target item, etc., which are not listed herein.

Similarly, the principle of generating cost floating information based on the item feature of the target item, or the principle of generating cost floating information in combination with the type information item feature of the target item and the like may be referred to in the above example, and details are not described here.

In some embodiments, S201 may be replaced with: and determining a preset average value as cost floating information, wherein the average value is generated based on the attribute information of at least one other article.

Exemplarily, if the target item is an item produced for the first time, and it is possible that the brand information and the category information of the target item are both on-line for the first time, the average value may be determined as the cost float information.

It should be noted that the average value may be generated for the attribute information of the at least one other item input based on the conversion cost model, or may be an average value of floating information of the cost of the at least one other item determined in other manners (for example, the page display device is configured based on the demand, history, experiment, and the like). If the mean value is generated based on the conversion cost model, the principle of generating the mean value based on the conversion generation model may be referred to in the above example, and the principle of generating the cost floating information based on the first attribute information of the target item is not described herein again.

S202: and generating conversion cost according to the cost floating information and the preset basic cost information.

For example, the basic cost information may be used to represent the cost information of the benchmark, and may be set by the page display device based on the requirement, history, experiment, and the like, which is not limited in this embodiment.

In some embodiments, the page display device may set the basic cost information corresponding to each brand by using the brand as a setting unit, may set the basic cost information corresponding to each category by using the category as a setting unit, may set the basic cost information corresponding to each region by using the region as a setting unit, and so on, which are not listed here.

Exemplarily, the conversion cost is cost float information ± base cost information.

S203: and acquiring second attribute information of the target object.

Note that the second attribute information may be the same attribute information as the first attribute information or may be different attribute information from the first attribute information.

For example, the first attribute information may be information related to a bias toward the target item itself, such as brand information and type information described in the above example, and the like, and the second attribute information may be information related to a bias toward feedback of the user with respect to the target item, such as a click and evaluation of the user, and the like.

Specifically, the second attribute information may include: text characteristics (such as search word text, search word subjects, search word brands and the like), article characteristics (such as article prices, article store information, article popularity number, article sales information and the like), user characteristics (such as user age, user gender, user historical behaviors, user evaluation sensitivity and the like), time characteristics (such as user access time, user access duration and the like), advertisement characteristics (such as advertiser delivery types, advertisement site flow information and the like) and Application (APP) characteristics (such as operating equipment information, regional information and the like).

S204: and inputting the second attribute information into the conversion rate model for prediction to generate the conversion rate.

In some embodiments, a method of constructing a conversion model may comprise:

step 1: a sample article is obtained.

Similarly, sample items are used for characterization, for training the items that generate the conversion rate model, and are not to be construed as limitations on the contents of the items. In addition, the selection of the type and the number of the sample articles is not limited in this embodiment, and the page display device may be set based on the requirement, the history, the experiment, and the like.

Step 2: and acquiring attribute information of the sample article.

Similarly, the attribute information of the sample article can be used for characterization, and is used for describing relevant information of the sample article, such as picture information, text information, user information, and the like.

And step 3: and scattering the attribute information of the sample object according to a preset deep cross network model and at least one training time sequence to obtain the attribute information of the scattered sample object.

For example, if a training time sequence is determined in units of weeks for a month, the training time sequence may be divided into 4 training time sequences, and attribute information of the sample article may be scattered according to time information in each training time sequence.

That is, the attribute information of the sample articles in one month may be divided into 4 groups in units of weeks (i.e., 7 days), each group corresponds to the attribute information of the sample articles in a corresponding time period, and the attribute information of the sample articles in the group is broken up for each group.

And 4, step 4: and acquiring a training data path from the attribute information of the scattered sample article based on a remainder mode.

The remainder taking manner is not limited in this embodiment, and may be, for example, modulo three remainder taking, and the like.

It should be understood that, the deep cross network model includes a plurality of nodes, each node may be used to assist in the relevant calculation of the prediction, and in this embodiment, each node may obtain a corresponding training data path from the attribute information of the broken sample article.

And 5: and generating a conversion rate model according to the deep cross network model and the training data path.

Illustratively, step 5 may include: determining sample characteristic vectors of attribute information of each sample object in each training data path according to a deep cross network model, training the sample characteristic vectors by combining parameters (such as weight coefficients, channel quantity and the like) of the deep cross network model, generating a prediction result (conversion rate obtained by training), comparing the prediction result with a preset prediction calibration result (conversion rate obtained by pre-calibration), generating a prediction loss function of the prediction result and the prediction calibration result, adjusting the deep cross network model according to the prediction loss function, specifically adjusting the parameters of the deep cross network model until the prediction loss function meets a preset prediction threshold or the prediction iteration number reaches a preset number threshold, and determining the corresponding deep cross network model as a conversion rate model.

Wherein, the training time sequence can be determined in a fixed time period, such as the above example, in units of weeks; or may be determined by a time period of change, such as a training time sequence every 7 days, a training time sequence every 4 days, a training time sequence every day, etc.; the determination may also be performed based on the attribute information of the sample article at different time periods, such as 150 days of the sample article and 30 days of the sample article.

It should be noted that, in this embodiment, the attribute information of the sample object is broken up in a training time sequence-based manner, and is trained to generate the conversion rate model, so that on one hand, the flexibility and diversity of the conversion rate model are realized, and the difficulty of breaking up the global information can be reduced due to the adoption of the breaking up in the training time sequence manner, thereby improving the training efficiency and saving the technical effect of training time.

In some embodiments, S204 may include:

step 1: and performing scattering operation on the second attribute information based on the conversion rate model and at least one preset time sequence.

For an exemplary principle of this step, the principle of the scattering operation performed on the attribute information of the sample item in the above example can be referred to, and details are not described here.

It should be noted that the sequence time sequence in the training process may be the same as or different from the time sequence in the prediction process, and this embodiment is not limited.

Step 2: and performing remainder calculation on the scattered second attribute information based on the conversion rate model to obtain a predicted data path.

For an exemplary principle of this step, the principle of obtaining the predicted data path in the above example may be referred to, and is not described herein again.

And step 3: and predicting the predicted data path based on the conversion rate model to generate the conversion rate.

S205: and generating bidding information according to the conversion cost and the conversion rate.

Exemplarily, the bid information is the cost of conversion and the conversion rate.

Based on the analysis, the conversion cost is generated based on cost floating information and basic cost information, the cost floating information can be a floating interval, the conversion cost is also a cost interval, and the bidding information can be a corresponding interval; of course, a cost floating value may be selected from the cost floating information, and accordingly, if the conversion cost is a specific cost value, the bidding information is also a specific bidding value, and the cost floating value may be selected as a maximum value of the floating interval or other values, which is not limited in this embodiment.

Illustratively, to provide accuracy and reliability of the bidding information, in some embodiments, adjustment parameters may be introduced, and S205 may be replaced with:

step 1: and generating an adjusting parameter according to the consumption information of the target item in the preset time period.

Illustratively, the consumption information may include: predicted consumption information and actual consumption information.

The predicted message information may be consumption information generated by the page display device based on the historical consumption record of the target item.

Illustratively, the time period is preferably the time period closest to the current time, such as a time within 24 hours of the closest current time.

Of course, in some embodiments, the adjustment parameter may also be set in units of time or in units of days, so as to improve flexibility of setting the adjustment parameter, thereby achieving the technical effect of determining reliability and accuracy of the bidding message.

Of course, in other embodiments, the adjustment parameters may be set by the page display device based on requirements, history, and tests.

Step 2: and generating bidding information according to the conversion cost, the conversion rate and the adjusting parameters.

Illustratively, bid information is conversion cost conversion rate adjustment parameter.

Fig. 4 is a diagram illustrating a page display method according to a third embodiment of the present application, as shown in fig. 4, the page display method includes:

s301: and the terminal equipment generates bidding information of each target item according to the conversion cost and the conversion rate.

For an exemplary description of generating bid information in S301, reference may be made to the description in the first embodiment or the second embodiment; and with respect to the description of obtaining the conversion cost and the conversion rate, reference may also be made to the description in the first embodiment or the second embodiment.

The terminal device may be, for example, a device of a vendor or a merchant that provides the output page with the target item.

S302: the terminal device sends bidding information to the electronic device.

For example, the electronic device may be a platform providing a display of a target item, such as an online shopping platform, or a live platform, and the target item is provided by a manufacturer or a merchant.

S303: and the electronic equipment ranks the target items according to the bidding information.

Illustratively, the electronic device may sort the target items in an ascending order based on the bid information or sort the target items in a descending order based on the bid information.

S304: the electronic device outputs a page including the target item based on the ranking result.

Illustratively, for the ascending sequence, the advertisement information to be displayed which is the last ordered in the ordering result is displayed in front of other advertisement information, and so on.

Illustratively, for the descending sequence, the advertisement information to be displayed which is ranked the first in the ranking result is displayed in front of other advertisement information, and so on.

It should be noted that the bid information and the output page may be determined by different execution bodies, such as the terminal-based device in the above embodiment, and the electronic device outputs the page according to the bid information. Of course, the bid information and output page may be determined by the same executing entity, such as by an electronic device and the page output based on the bid information.

Fig. 5 is a fourth embodiment according to the present application, and as shown in fig. 5, the page display method of the embodiment includes:

s401: obtaining bidding information of each target item to be displayed, wherein the bidding information is generated based on the conversion cost and the conversion rate.

For example, the execution main body of the embodiment may be a page display device, and the page display device may also be a server (including a local server and a cloud server), a terminal device, a processor, a chip, and the like, which is not limited in this embodiment.

For example, when the page display method of the present embodiment is applied to an application scenario as shown in fig. 1, the page display apparatus may be a terminal device.

It should be noted that the conversion cost and the conversion rate may be generated for the execution subject in this embodiment, or may be obtained from an external device, such as the terminal device in the third embodiment.

S402: and ordering the target items according to the bidding information.

S403: and outputting a page comprising the target item based on the sorting result.

In some embodiments, the cost float information is generated by inputting the first attribute information of the target item into the conversion cost model for prediction; or the cost floating information is a preset average value, wherein the preset average value is generated by inputting the attribute information of at least one other article into the conversion cost model for prediction.

In some embodiments, if the first attribute information includes brand information and/or category information, the cost float information is generated by converting a cost model to predict a brand feature vector and/or a category feature vector;

the brand feature vector is generated by the conversion cost model based on brand information, and the category feature vector is generated by the conversion cost model based on category information.

In some embodiments, the conversion rate is generated by inputting the second attribute information of the target object into a preset conversion rate model for prediction.

In some embodiments, the conversion rate is obtained by predicting the second property information after scattering based on a conversion rate model, and the second property information after scattering is obtained by scattering the second property information in at least one preset time sequence based on the conversion rate model.

In some embodiments, the conversion rate is predicted from a predicted data path based on a conversion rate model, and the predicted data path is calculated by taking a remainder from the broken second attribute information.

In some embodiments, the bid information is generated based on conversion costs, conversion rates, and preset tuning parameters;

In some embodiments, the conversion cost model is obtained by inputting the attribute information of the object to be trained into the feedforward neural network model for training.

Fig. 6 is a fifth embodiment of the present application, and as shown in fig. 6, the method for training a model applied to page display of the present embodiment includes:

s501: and acquiring attribute information of the sample article.

For example, the execution subject of this embodiment may be a model training device, and the model training device may be a server (including a local server and a cloud server), a terminal device, a processor, a chip, and the like, which is not limited in this embodiment.

It should be noted that the execution main body of the present embodiment may be the same as the execution main body of the first embodiment, may also be the same as the execution main body of the fourth embodiment, and may also be different from the execution main body of the first embodiment and the execution main body of the fourth embodiment, and the present embodiment is not limited.

S502: and scattering the attribute information of the sample object according to a preset deep cross network model and at least one preset training time sequence to obtain the attribute information of the scattered sample object.

S503: and training the attribute information of the scattered sample articles according to the deep cross network model to obtain a conversion rate model.

The conversion rate model is used for generating conversion rate, the conversion rate and the conversion cost are used for generating bidding information of each target item, and the bidding information is used for obtaining the ranking result of the target items.

In some embodiments, S503 may include:

step 1: and determining a training data path according to the attribute information of the scattered sample articles.

In some embodiments, the step may specifically include: and acquiring a training data path from the sample attribute information of the scattered sample articles based on a remainder mode.

Step 2: and training according to the training data path, and generating and outputting a conversion cost model.

In some embodiments, S503 may include:

step 1: and determining a sample characteristic vector of the attribute information of the scattered sample article.

Step 2: a sample prediction result corresponding to the sample feature vector is generated.

And step 3: and adjusting the deep cross network model according to the sample prediction result to generate a conversion rate model.

In some embodiments, the step may specifically include:

step 1: and generating a sample loss function of the sample prediction result and a preset sample calibration result.

Step 2: and adjusting the cross network model according to the sample loss function, and generating and outputting a conversion cost model.

Fig. 7 is a sixth embodiment according to the present application, and as shown in fig. 7, a page display apparatus of the present embodiment is configured to execute the method according to the first embodiment or the second embodiment, and includes:

the generating module 11 is configured to generate bidding information of each target item according to the conversion cost and the conversion rate;

the sending module 12 is configured to send the bid information of each target item to an electronic device, where the bid information of each target item is used to rank the target items to obtain a ranking result, and the ranking result is used to output a page including the target items.

Fig. 8 is a seventh embodiment according to the present application, and as shown in fig. 8, the page display apparatus of the present embodiment further includes:

a first obtaining module 13, configured to obtain first attribute information of the target item, input the first attribute information to the conversion cost model for prediction, and generate the cost floating information; alternatively, the first and second electrodes may be,

the first obtaining module 13 is configured to obtain attribute information of at least one other article, input the attribute information of the other article to the conversion cost model for prediction, generate a mean value, and use the mean value as the cost floating information.

In some embodiments, the first obtaining module 13 is configured to obtain a brand feature vector of the brand information, and/or obtain a category feature vector of the category information; and predicting according to the brand feature vector and/or the category feature vector to generate the cost floating information.

In some embodiments, the first obtaining module 13 is configured to obtain second attribute information of the target item, and input the second attribute information to a conversion rate model for prediction, so as to generate the conversion rate.

In some embodiments, the first obtaining module 13 is configured to break up the second attribute information according to at least one preset time sequence, predict the broken second attribute information, and generate the conversion rate.

In some embodiments, the first obtaining module 13 is configured to perform remainder calculation on the second attribute information after being scattered to obtain a predicted data path, predict the predicted data path, and generate the conversion rate.

In some embodiments, the generating module 11 is configured to generate an adjustment parameter according to consumption information of the target item within a preset time period, and generate the bidding information according to the conversion cost, the conversion rate, and the adjustment parameter.

Fig. 9 is a diagram illustrating an eighth embodiment of the present application, where as shown in fig. 9, a page display apparatus of the present embodiment is configured to perform the method according to the fourth embodiment, and includes:

a second obtaining module 21, configured to obtain bid information of each target item to be displayed, where the bid information is generated based on a conversion cost and a conversion rate;

a sorting module 22, configured to sort the target items according to the bidding information;

and the output module 23 is used for outputting the page comprising the target item based on the sorting result.

Fig. 10 is a diagram illustrating a model training apparatus applied to a page display according to a ninth embodiment of the present application, as shown in fig. 10, for executing the method according to the fifth embodiment, including:

a third obtaining module 31, configured to obtain attribute information of the sample article;

a scattering module 32, configured to scatter the attribute information of the sample object according to a preset deep cross network model and at least one preset training time sequence, so as to obtain the attribute information of the scattered sample object;

the training module 33 is configured to train the attribute information of the scattered sample articles according to the deep cross network model to obtain a conversion rate model; the conversion rate model is used for generating a conversion rate, the conversion rate and the conversion cost are used for generating bidding information of each target item, and the bidding information is used for obtaining a sequencing result of the target items.

In some embodiments, the scattering module 32 is configured to determine a training data path according to the attribute information of the scattered sample article, perform training according to the training data path, and generate and output the conversion cost model.

In some embodiments, the breaking module 32 is configured to obtain the training data path from the sample attribute information of the broken sample item based on a remainder manner.

In some embodiments, the training module 33 is configured to determine a sample feature vector of the attribute information of the scattered sample article, generate a sample prediction result corresponding to the sample feature vector, and adjust the deep cross network model according to the sample prediction result to generate the conversion rate model.

In some embodiments, the training module 33 is configured to generate a sample loss function of the sample prediction result and a preset sample calibration result, adjust the cross network model according to the sample loss function, and generate and output the conversion cost model.

An embodiment of the present application further provides a page display system, configured to perform the method according to any one of the first to fifth embodiments, including: the apparatus of the sixth embodiment or the apparatus of the seventh embodiment, the apparatus of the eighth embodiment.

In some embodiments, the system further comprises the apparatus of the ninth embodiment.

FIG. 11 is a block diagram of a tenth embodiment of the present application, wherein the computer device of the present embodiment is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers, as shown in FIG. 11. The computer device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.

As shown in fig. 11, the computer apparatus includes: one or more processors 101, memory 102, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. 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 for execution within the computer device, including instructions stored in or on the memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to the interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple computer devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). Fig. 11 illustrates an example of one processor 101.

Memory 102 is a non-transitory computer readable storage medium as provided herein. The memory stores instructions executable by at least one processor to cause the at least one processor to perform the page display method or the model training method applied to page display provided by the application. The non-transitory computer-readable storage medium of the present application stores computer instructions for causing a computer to execute the page display method or the model training method applied to page display provided by the present application.

Memory 102, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and modules. The processor 101 executes various functional applications of the server and data processing, i.e., implementing the page display method or the model training method applied to the page display in the above-described method embodiments, by running non-transitory software programs, instructions, and modules stored in the memory 102.

The memory 102 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of the computer device, and the like. Further, the memory 102 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 102 may optionally include memory located remotely from processor 101, which may be connected to a computer 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 computer device may further include: an input device 103 and an output device 104. The processor 101, the memory 102, the input device 103, and the output device 104 may be connected by a bus or other means, and fig. 11 illustrates an example of connection by a bus.

The input device 103 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the computer apparatus, such as a touch screen, keypad, mouse, track pad, touch pad, pointer stick, one or more mouse buttons, track ball, joystick, or other input device. The output devices 104 may include a display device, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibrating 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 can be a touch screen.

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

These computer programs (also known as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. 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 a pointing device (e.g., a mouse or a 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 can 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, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end 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 back-end, 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 clients and servers. A client and server are generally 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.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method of page display, the method comprising:

2. The method of claim 1, wherein the conversion cost is derived based on preset base cost information and cost float information.

3. The method of claim 2, wherein the method further comprises:

4. The method of claim 3, wherein the first attribute information includes brand information and/or category information; inputting the first attribute information into a preset conversion cost model for prediction to generate the cost floating information;

5. The method of claim 1, wherein the method further comprises:

acquiring second attribute information of the target object;

6. The method of claim 5, wherein inputting the second attribute information into a conversion model for prediction, generating the conversion, comprises:

7. The method of claim 6, wherein predicting the broken-up second attribute information to generate the conversion rate comprises:

and predicting the predicted data path to generate the conversion rate.

8. The method of any one of claims 1 to 7, wherein generating bid information for each target item based on conversion cost and conversion rate comprises:

9. The method of claim 8, wherein the consumption information comprises predicted consumption information and actual consumption information.

10. The method according to any one of claims 3-7, wherein the conversion cost model is obtained by inputting attribute information of the article to be trained into a feed-forward neural network model for training.

11. A method of page display, the method comprising:

sorting the target items according to the bidding information;

outputting a page including the target item based on the sorting result.

12. The method of claim 11, wherein the conversion cost is derived based on pre-set base cost information and cost float information.

13. The method of claim 12, wherein the cost float information is generated from input of first attribute information of the target item into a conversion cost model for prediction; or the cost floating information is a preset average value, wherein the preset average value is generated by inputting the attribute information of at least one other article to the conversion cost model for prediction.

14. The method of claim 13, wherein if the first attribute information includes brand information and/or category information, the cost float information is generated by the translation cost model predicting a brand feature vector and/or a category feature vector;

15. The method of claim 11, wherein the conversion rate is predicted by inputting the second attribute information of the target item to a preset conversion rate model.

16. The method of claim 15, wherein the conversion rate is predicted based on the conversion rate model, and the broken second attribute information is obtained by breaking the second attribute information in at least one preset time sequence based on the conversion rate model.

17. The method of claim 16, wherein the conversion rate is predicted from a predicted data path based on the conversion rate model, the predicted data path being calculated by taking a remainder from the broken second attribute information.

18. The method of any of claims 11 to 17, wherein the bid information is generated based on the conversion cost, the conversion rate, and preset tuning parameters;

19. The method of claim 18, wherein the consumption information comprises predicted consumption information and actual consumption information.

20. The method according to any one of claims 13-17, wherein the conversion cost model is obtained by inputting attribute information of an article to be trained into a feed-forward neural network model for training.

21. A model training method applied to page display, the method comprising:

acquiring attribute information of a sample article;

22. The method of claim 21, wherein training the attribute information of the broken sample items according to the deep cross-network model to obtain a conversion rate model comprises:

23. The method of claim 22, wherein determining a training data path from attribute information of the broken sample item comprises: and obtaining the training data path from the sample attribute information of the scattered sample articles based on a remainder mode.

24. The method of any one of claims 21 to 23, wherein training the attribute information of the broken sample items according to the deep cross-network model to obtain a conversion rate model comprises:

25. The method of claim 24, wherein adjusting the deep cross network model based on the sample predictions to generate the conversion model comprises:

26. A page display apparatus comprising:

27. A page display apparatus comprising:

28. A model training device applied to page display comprises:

29. A page display system, comprising:

the apparatus of claim 26;

the apparatus of claim 27.

30. A computer device, comprising: a memory, a processor;

a memory; a memory for storing the processor-executable instructions;

wherein the processor is configured to perform the method of any one of claims 1 to 10; alternatively, the first and second electrodes may be,

the processor is configured to perform the method of any one of claims 11 to 20; alternatively, the first and second electrodes may be,

the processor is configured to perform the method of any one of claims 21 to 25.

31. A computer readable storage medium having stored therein computer executable instructions for implementing the method of any one of claims 1 to 10 when executed by a processor; alternatively, the first and second electrodes may be,

the computer executable instructions when executed by a processor are for implementing the method of any one of claims 11 to 20; alternatively, the first and second electrodes may be,

the computer executable instructions when executed by a processor are for implementing the method of any one of claims 21 to 25.