CN111582944A

CN111582944A - Crowd expansion method, device, equipment and storage medium for advertisement

Info

Publication number: CN111582944A
Application number: CN202010408433.6A
Authority: CN
Inventors: 王玉昕; 郑祺星
Original assignee: Beijing Kingsoft Internet Security Software Co Ltd
Current assignee: Beijing Kingsoft Internet Security Software Co Ltd
Priority date: 2020-05-14
Filing date: 2020-05-14
Publication date: 2020-08-25
Anticipated expiration: 2040-05-14
Also published as: CN111582944B

Abstract

According to the crowd expansion method, the device, the equipment and the storage medium for the advertisement, the first user information of the seed user of a first advertiser and the second user information of the seed user of a second advertiser are obtained; the first advertiser is an advertiser of the advertisement to be promoted, and the second advertiser is an advertiser different from the first advertiser; performing crowd expansion respectively based on the first user information and the second user information to obtain third user information of a target user corresponding to the first user information and fourth user information of the target user corresponding to the second user information; obtaining seed users of the advertisements to be promoted based on the third user information and the fourth user information; and carrying out crowd expansion on the basis of the seed users of the advertisements to be promoted to obtain target users of the advertisements to be promoted. The scheme can improve the richness of the target users of the advertising crowd expansion.

Description

Crowd expansion method, device, equipment and storage medium for advertisement

Technical Field

The invention relates to the technical field of crowd expansion, in particular to a crowd expansion method, a device, equipment and a storage medium for advertisements.

Background

In order to reduce the problem of high advertisement promotion cost caused by promotion to all users, the crowd expansion can be carried out aiming at the advertisement to be promoted: and finding a certain number of similar crowds with relevance as target users of the advertisement to be promoted. Specifically, based on the seed user, through a certain algorithm evaluation model, a plurality of users having relevance with the seed user are found from a user resource library and serve as target users of the advertisement to be promoted.

However, in the process of implementing the present invention, the inventor finds that the seed user is often submitted by the advertiser, and is influenced by factors such as limited user resources owned by the advertiser, the seed user is relatively single, and further the target user richness determined based on the seed user is low.

Disclosure of Invention

The embodiment of the invention aims to provide an expansion method, an expansion device, expansion equipment and a storage medium, so as to achieve the effect of improving the richness of target users obtained by expanding advertising crowds. The specific technical scheme is as follows:

in a first aspect, an embodiment of the present invention provides a crowd expansion method, including:

acquiring first user information of a seed user of a first advertiser and acquiring second user information of a seed user of a second advertiser; the first advertiser is an advertiser of the advertisement to be promoted, and the second advertiser is an advertiser different from the first advertiser;

performing crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information, and performing crowd expansion based on the second user information to obtain fourth user information of the target user corresponding to the second user information;

obtaining seed users of the advertisement to be promoted based on the third user information and the fourth user information;

and carrying out crowd expansion based on the seed users of the advertisements to be promoted to obtain target users of the advertisements to be promoted.

Optionally, the obtaining first user information of the seed user of the first advertiser includes:

acquiring user information of a first number of users who perform appointed operation on the product of the first advertiser to obtain first positive user information; wherein the specified operation is an operation on advertisement conversion of the product of the advertiser;

acquiring user information of a first number of users who do not perform designated operation on the product of the first advertiser to obtain first negative user information;

and taking the first positive user information and the first negative user information as first user information of seed users of the first advertiser.

Optionally, the crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information includes:

acquiring information on an advertisement conversion result of a product of an advertiser and information which can influence the advertisement conversion result of the product of the advertiser from the first positive user information and the first negative user information of the first user information;

constructing a first user characteristic of a seed user of the first advertiser using the obtained information;

training to obtain a first model for obtaining a first classification result of the user by using the first user characteristic; wherein the first classification result comprises: a first positive target type having a user characteristic indicated by the first positive user information or a first negative target type having a user characteristic indicated by the first negative user information;

inputting the user information in a pre-stored user resource library into the first model to obtain a first classification result of the user information in the pre-stored user resource library;

and selecting the obtained first classification result as the user information of the first positive target type as the third user information of the target user corresponding to the first user information.

Optionally, the obtaining second user information of the seed user of the second advertiser includes:

acquiring user information of a user who performs the specified operation on the product of the second advertiser to obtain second positive user information;

acquiring user information of users who successfully popularize and do not perform the specified operation from users who are popularization objects of the product of the second advertiser to obtain second negative user information;

and taking the second positive user information and the second negative user information as second user information of seed users of the second advertiser.

Optionally, the crowd expansion based on the second user information to obtain fourth user information of a target user corresponding to the second user information includes:

acquiring information on an advertisement conversion result of a product of an advertiser and information which can influence the advertisement conversion result of the product of the advertiser from the second positive user information and the second negative user information of the second user information;

constructing a second user characteristic of a seed user of the second advertiser using the obtained information;

training to obtain a second model for obtaining a second classification result of the user by utilizing the second user characteristic; wherein the second classification result comprises: a second positive target type having a user characteristic indicated by the second positive user information or a second negative target type having a user characteristic indicated by the second negative user information;

inputting the user information in the pre-stored user resource library into the second model to obtain a second classification result of the user information in the pre-stored user resource library;

and selecting the user information of which the obtained second classification result is the second positive target type as fourth user information of a target user corresponding to the second user information.

Optionally, the obtaining the seed user of the advertisement to be promoted based on the third user information and the fourth user information includes:

promoting the advertisement to be promoted to the initial user indicated by the third user information and the initial user indicated by the fourth user information, and acquiring operation data of each initial user on the advertisement to be promoted;

and selecting users meeting preset promotion effect conditions from all initial users as seed users of the advertisements to be promoted based on the operation data.

Optionally, the operation data includes: data indicating whether the advertisement to be promoted is successfully promoted or not and data indicating whether the specified operation is performed on the advertisement to be promoted or not;

the selecting, based on the operation data, a user that meets a preset promotion effect condition from among the initial users as a seed user of the advertisement to be promoted includes:

when the operation data indicate that the number of the users who successfully popularize the advertisement to be popularized is equal to a first number in the initial users, acquiring user information of the user who performs the specified operation in the initial users to obtain third positive user information;

acquiring user information of users who succeed in popularizing the advertisement to be promoted and do not perform the specified operation in each initial user to obtain third negative user information;

and taking the users corresponding to the third positive user information and the third negative user information as seed users of the advertisement to be promoted.

Optionally, the crowd expansion is performed on the seed users based on the advertisement to be promoted to obtain target users of the advertisement to be promoted, including:

acquiring information about an advertisement conversion result of the advertisement to be promoted and information which can influence the advertisement conversion result of the advertisement to be promoted in the third positive user information and the third negative user information of the seed user of the advertisement to be promoted;

constructing a third user characteristic of the seed user of the advertisement to be promoted by using the acquired information;

training to obtain a third model for obtaining a third classification result of the user by using the third user characteristic; wherein the third classification result comprises: a third positive target type having a user characteristic indicated by the third positive user information or a third negative target type having a user characteristic indicated by the third negative user information;

inputting the user information in the pre-stored user resource library into the third model to obtain a third classification result of the user information in the pre-stored user resource library;

and selecting the user corresponding to the user information of which the obtained third classification result is the third positive target type as the target user of the advertisement to be promoted.

In a second aspect, an embodiment of the present invention provides a crowd expansion apparatus for advertisement, including:

the system comprises a user information acquisition module, a first advertisement server and a second advertisement server, wherein the user information acquisition module is used for acquiring first user information of seed users of a first advertiser and acquiring second user information of seed users of a second advertiser; the first advertiser is an advertiser of the advertisement to be promoted, and the second advertiser is an advertiser different from the first advertiser;

the initial user acquisition module is used for carrying out crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information and carrying out crowd expansion based on the second user information to obtain fourth user information of the target user corresponding to the second user information;

a seed user obtaining module, configured to obtain a seed user of the advertisement to be promoted based on the third user information and the fourth user information; (ii) a

And the target user expansion module is used for expanding the crowd based on the seed users of the advertisements to be promoted to obtain the target users of the advertisements to be promoted.

Optionally, the user information obtaining module is specifically configured to:

Optionally, the initial user acquiring module is specifically configured to:

Optionally, the seed user obtaining module is specifically configured to:

promoting the advertisement to be promoted to the initial user indicated by the third user information and the initial user indicated by the fourth user information, and acquiring operation data of each initial user on the advertisement to be promoted; and selecting users meeting preset promotion effect conditions from all initial users as seed users of the advertisements to be promoted based on the operation data.

the seed user acquisition module is specifically configured to:

Optionally, the target user expansion module includes:

the characteristic construction submodule is used for acquiring information about an advertisement conversion result of the advertisement to be promoted and information which can influence the advertisement conversion result of the advertisement to be promoted in the third positive user information and the third negative user information of the seed user of the advertisement to be promoted; constructing a third user characteristic of the seed user of the advertisement to be promoted by using the acquired information;

the model training submodule is used for training to obtain a third model for obtaining a third classification result of the user by utilizing the third user characteristic; wherein the third classification result comprises: a third positive target type having a user characteristic indicated by the third positive user information or a third negative target type having a user characteristic indicated by the third negative user information;

the user expansion submodule is used for inputting the user information in a pre-stored user resource library into the third model to obtain a third classification result of the user information in the pre-stored user resource library; and selecting the user corresponding to the user information of which the obtained third classification result is the third positive target type as the target user of the advertisement to be promoted.

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

the system comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete mutual communication through the bus; a memory for storing a computer program; and the processor is used for executing the program stored in the memory and realizing the steps of the advertisement crowd expansion method provided by the first aspect.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where a computer program is stored in the storage medium, and when the computer program is executed by a processor, the steps of the crowd expansion method for advertisement provided in the first aspect are implemented.

In the scheme provided by the embodiment of the invention, the first advertiser is an advertiser of the advertisement to be promoted, and the second advertiser is an advertiser different from the first advertiser. Therefore, the first user information of the seed user of the first advertiser and the second user information of the seed user of the second advertiser are obtained, crowd expansion is further carried out based on the first user information, third user information of a target user corresponding to the first user information is obtained, crowd expansion is carried out based on the second user information, and fourth user information of the target user corresponding to the second user information is obtained. On the basis of the information, the seed users of the advertisements to be promoted are obtained based on the third user information and the fourth user information, and compared with the seed users provided by the first advertiser for the advertisements to be promoted, the seed users are more diverse, so that the richness of the target users obtained by crowd expansion of the seed users based on the advertisements to be promoted can be improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.

Fig. 1 is a flowchart illustrating a crowd expansion method for advertisement according to an embodiment of the present invention;

FIG. 2 is a flowchart illustrating a crowd expansion method for advertisement according to another embodiment of the present invention;

FIG. 3 is a diagram illustrating an exemplary configuration of a crowd expansion apparatus for advertisement according to an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of a crowd expansion apparatus for advertisement according to another embodiment of the present invention;

fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

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

First, a method for expanding a population of an advertisement according to an embodiment of the present invention will be described.

The crowd expansion method for the advertisement provided by the embodiment of the invention can be applied to electronic equipment, and the equipment can specifically include a desktop computer, a portable computer, an internet television, an intelligent mobile terminal, a server, a wearable intelligent terminal and the like, and is not limited herein, and any electronic equipment capable of implementing the embodiment of the invention belongs to the protection scope of the embodiment of the invention.

As shown in fig. 1, a flow of a crowd expansion method for advertisement according to an embodiment of the present invention may include the following steps:

s101, first user information of a seed user of a first advertiser is obtained, and second user information of a seed user of a second advertiser is obtained.

The first advertiser is an advertiser of the advertisement to be promoted, and the second advertiser is an advertiser different from the first advertiser.

In order to improve the richness of the target users obtained based on the subsequent seed user expansion, besides the first user information of the seed user of the first advertiser, the second user information of the seed user of a second advertiser different from the first advertiser can be obtained, and different seed users are added through different advertisers, so that the diversity of the seed users is improved. In a specific application, the manner of obtaining the first user information of the seed user of the first advertiser and obtaining the second user information of the seed user of the second advertiser may be various. For example, first user information of a seed user sent by a first advertiser may be received, and similarly, second user information of a seed user sent by a second advertiser may be received. Or, for example, a first sub-user selection condition sent by a first advertiser may be received, and user information meeting the first sub-user selection condition is selected from a user resource library as first user information; similarly, a second seed user selection condition sent by a second advertiser may be received, and user information satisfying the second seed user selection condition may be selected from the user resource library as second user information.

Or, for example, different user information may be acquired as to whether the user performs an operation regarding advertisement conversion of the advertiser's product. The advertisement conversion refers to an operation of converting the identity of the user from a pushing object of the advertisement to a using object of a product in the advertisement. For example, when the product in the advertisement is a website, the advertisement conversion may include: browse the website, or become a registered user of the website, and the like. When the product in the advertisement is an APP (Application), the advertisement conversion may include: install APP or activate APP, etc. When the product in the advertisement is an item sold by the e-commerce platform, the advertisement conversion may include: collecting the item or purchasing the item, etc. The third exemplary description is followed by a detailed description in the form of an alternative embodiment for ease of understanding and reasonable layout. Any method capable of obtaining the first user information and the second user information may be used in the present invention, and this embodiment does not limit this.

Also, any user information may be various. For example, the user information may be a personalized tag of the user, such as an age tag, a gender tag, and a city tag of the user, and so on. Alternatively, the user information may be, for example, information that affects the user's preference for the advertiser's product. For example, when the product of the advertiser is APP, the user information may include information on a name of APP installed by the user, a name of APP uninstalled by the user, an operating system of the electronic device used by the user, and a brand of the electronic device, and the like. When the advertiser's product is an item sold by an e-commerce, the user information may include information on items viewed by the user on the e-commerce platform, favorite items, uninteresting items, and purchased items, among others.

S102, carrying out crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information, and carrying out crowd expansion based on the second user information to obtain fourth user information of the target user corresponding to the second user information.

In a specific application, the crowd expansion mode based on the user information of the seed user can be various. For example, when the personalized tag exists in the user information of the seed user, the user information of the user having the same personalized tag as the seed user may be selected from the user resource library, so as to obtain the user information of the target user. Or, for example, the user information of the seed user may be classified according to whether the seed user is interested in the product of the advertiser, the user information of the seed user interested in the product of the advertiser is used as positive user information, and the user information of the seed user not interested in the product of the advertiser is used as negative user information; and training by utilizing the positive user information and the negative user information to obtain a classification model, classifying the users in the user resource library by utilizing the classification model obtained by training, and taking the users with the types of the positive user information as the user information of the target user. Or, for example, when the seed user has a friend relationship of the social network, the user information of the friend of the seed user may be used as the user information of the target user.

When crowd expansion is performed based on the first user information to obtain third user information of a target user corresponding to the first user information and crowd expansion is performed based on the second user information to obtain fourth user information of the target user corresponding to the second user information, the crowd expansion mode in any one of the above exemplary descriptions can be adopted, and the difference lies in that the user information of the seed user can be replaced by the first user information or the second user information in a targeted manner. The second exemplary description is followed by a detailed description in the form of an alternative embodiment for ease of understanding and reasonable layout.

S103, obtaining seed users of the advertisements to be promoted based on the third user information and the fourth user information.

In a specific application, the seed users obtaining the advertisement to be promoted based on the third user information and the fourth user information may be various, and the following description is made in an alternative embodiment.

In an optional implementation manner, the obtaining of the seed user of the advertisement to be promoted based on the third user information and the fourth user information may specifically include the following steps:

and taking the user indicated by the third user information and the user indicated by the fourth user information as seed users of the advertisement to be promoted.

In another optional implementation manner, the obtaining of the seed user of the advertisement to be promoted based on the third user information and the fourth user information may specifically include the following steps:

promoting the advertisement to be promoted to the initial user indicated by the third user information and the initial user indicated by the fourth user information, and acquiring operation data of the advertisement to be promoted by each initial user;

and based on the operation data, selecting users meeting the preset promotion effect conditions from the initial users as seed users of the advertisements to be promoted.

In a specific application, if the diversity of seed users is expanded at will, the number of invalid users who do not have interest in the advertisement to be promoted in the expanded seed users may be too large, and further, target users determined based on the seed users may not be accurate enough. Therefore, in order to reduce users who are not interested in the advertisement to be promoted among the initial users and improve the accuracy of the subsequently determined target user, the advertisement to be promoted can be promoted for the initial user indicated by the third user information and the initial user indicated by the fourth user information, the operation data of the advertisement to be promoted of each initial user can be obtained, and then the user meeting the preset promotion effect condition is selected from the initial users based on the operation data and serves as the seed user of the advertisement to be promoted.

Moreover, the operation data of the advertisement to be promoted by each initial user can be obtained in various ways. For example, the operation data of the advertisement to be promoted by each initial user may be requested from the electronic device of the user that received the advertisement to be promoted. Or, for example, operation data of the advertisement to be promoted by each initial user may be requested from the electronic device corresponding to the advertisement to be promoted. For example, the electronic device corresponding to the advertisement to be promoted may be a server of a video APP for displaying the advertisement. The preset promotion effect condition is a condition for selecting a user interested in the advertisement to be promoted from all initial users. For the convenience of understanding and reasonable layout, the step S104 is described in detail in the embodiment of fig. 2 of the present invention.

In the optional embodiment, the advertisement to be promoted is promoted by the initial user indicated by the third user information and the initial user indicated by the fourth user information, and operation data of the advertisement to be promoted by each initial user is acquired; and then based on the operation data, selecting users meeting the preset promotion effect conditions from the initial users as seed users of the advertisements to be promoted, and screening to obtain users which are relatively interested in the advertisements to be promoted from the initial users as target users, so that the accuracy of the target users of the advertisements to be promoted is improved, and the richness and the accuracy of the target users obtained by expanding advertisement crowds are considered.

The operation data may specifically include: and the data indicates whether the advertisement to be promoted is successfully promoted and/or indicates whether the advertisement to be promoted is subjected to specified operation. For example, the data indicating whether the advertisement to be promoted is successfully promoted may include: whether the advertisement to be promoted is successfully displayed on the electronic equipment of the user or whether the advertisement to be promoted is successfully pushed to the electronic equipment of the user. The data indicating whether the advertisement to be promoted is subjected to the specified operation may include: when the advertisement to be promoted contains the APP, the advertisement can be data related to the user to install the APP, for example, data requesting a download link of the APP and requesting an installation package of the APP; alternatively, it may be data about user activation of the APP, such as data for registering the APP and a request for obtaining the authentication code.

And S104, performing crowd expansion based on the seed users of the advertisements to be promoted to obtain target users of the advertisements to be promoted.

The specific manner of performing crowd expansion based on the seed user of the advertisement to be promoted to obtain the target user of the advertisement to be promoted is similar to the manner of performing crowd expansion based on the user information of the seed user to obtain the target user corresponding to the user information of the seed user in step S102. The difference is that in step S104 the seed user is the seed user of the advertisement to be promoted. For easy understanding and reasonable layout, in the following, in an optional embodiment of fig. 2 of the present invention, a manner of performing crowd expansion on seed users based on an advertisement to be promoted to obtain target users of the advertisement to be promoted is specifically described.

In an optional implementation manner, the obtaining of the first user information of the seed user of the first advertiser may specifically include the following steps:

acquiring user information of a first number of users who perform appointed operation on a product of a first advertiser to obtain first positive user information; wherein the specified operation is an operation on advertisement conversion of the product of the advertiser;

acquiring user information of a first number of users who do not perform designated operation on a product of a first advertiser to obtain first negative user information;

and taking the first positive user information and the first negative user information as the first user information of the seed user of the first advertiser.

In particular applications, the specified actions regarding advertising conversions for the advertiser's products may be varied. For example, when the product of the advertiser is an APP, the specified operation may be that the APP of the advertiser is installed; alternatively, when the advertiser's product is an item sold on the e-commerce platform, the specified action may be the purchase of the item sold by the advertiser. Accordingly, the unspecified operation may be plural. For example, when the product of the advertiser is an APP, the unspecified operation may be that the APP of the advertiser has never been installed; alternatively, when the advertiser's product is an item sold on the e-commerce platform, the specified action may be that the advertiser's sold item has never been purchased.

Also, since the designated operation is an operation on advertisement conversion of the advertiser's product, the first positive user information may indicate information of users who are relatively more interested in the advertisement of the first advertiser, and the first negative user information may be information of users who are relatively not interested in the advertisement of the first advertiser. On the basis, the first positive user information and the first negative user information are used as the first user information of the seed user of the first advertiser, so that the subsequent expansion of users which are relatively easy to accept advertisements and are correspondingly more likely to be interested in the advertisements to be promoted from the user resource library is facilitated.

In an optional implementation manner, the crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information may specifically include the following steps:

acquiring information on an advertisement conversion result of a product of an advertiser and information which can influence the advertisement conversion result of the product of the advertiser from first positive user information and first negative user information of the first user information;

constructing a first user characteristic of a seed user of the first advertiser by using the acquired information;

training to obtain a first model for obtaining a first classification result of the user by utilizing the first user characteristic; wherein the first classification result comprises: a first positive target type having a user characteristic indicated by the first positive user information, or a first negative target type having a user characteristic indicated by the first negative user information;

inputting the pre-stored user information in the user resource library into a first model to obtain a first classification result of the pre-stored user information in the user resource library;

In a particular application, information regarding the advertisement conversion results of the advertiser's products, as well as information that can affect the advertisement conversion results of the advertiser's products, may be varied. For example, when the advertiser's product is an APP, the information on the advertisement conversion result of the advertiser's product may include: the user installs the names of the APPs in a specified time period, for example, in the last three months, unloads the names of the APPs, retains the names of the APPs, and uses the names of the APPs with the frequency greater than a preset threshold. Also, the information that can affect the advertisement conversion result of the advertiser's product may include: the electronic equipment of the user comprises information such as the brand, the size, the name of an operator and the resident province of the user of the intelligent mobile terminal.

Or, for example, when the advertiser's product is an item sold by the e-commerce platform, the information on the advertisement conversion result of the advertiser's product may include: the name of the item purchased by the user within a specified time period, such as the last three months, the name of the item marked as not interesting, the name of the collected item, and the name of the shared item. Also, the information that can affect the advertisement conversion result of the advertiser's product may include: the electronic device of the user is information such as the brand, size, operator name, resident province of the user, and the like of the smart mobile terminal.

And, the first user characteristic of the seed user of the first advertiser is constructed by using the obtained information, which may be obtained by splicing the obtained information to obtain the first user characteristic, or by forming an array of the obtained information to obtain the first user characteristic, and so on. When a first model for obtaining a first classification result of a user is obtained by training using a first user characteristic, in order to deal with the characteristic that the data volume of the first user characteristic of a seed user of a first advertiser is large, a logistic regression model can be selected as the trained model to improve the training efficiency, and then model parameters of the logistic regression model are adjusted through training to obtain the first model. In addition, in order to deal with the characteristic that the feature dimension of the first user feature is relatively high, the regular regression function can be added in an addition form in the cost function used in the training process of the logistic regression model, so that the problem that the first model obtained by training is over-fitted on a certain feature and only user information with the feature is selected in a concentrated mode is solved, and the accuracy of population expansion is further improved.

In an optional implementation manner, the obtaining of the second user information of the seed user of the second advertiser may specifically include the following steps:

acquiring user information of a user who performs specified operation on a product of a second advertiser to obtain second positive user information;

acquiring user information of users who are successful in popularization and do not perform specified operation from users who are popularization objects of products of a second advertiser to obtain second negative user information;

and taking the second positive user information and the second negative user information as second user information of the seed user of the second advertiser.

In particular applications, the specified actions regarding advertising conversions for the advertiser's products may be varied. For example, for the second advertiser, when the product of the advertiser is an APP, the specified action may be to install or activate the advertiser's APP; alternatively, when the advertiser's product is an item sold on the e-commerce platform, the specified action may be the purchase of the item sold by the advertiser. Accordingly, the unspecified operation may be plural. For example, when the product of the advertiser is an APP, the non-specified operation may be that the APP of the advertiser is not installed or activated; alternatively, when the advertiser's product is an item sold on the e-commerce platform, the specified action may be that the advertiser's sold item has never been purchased. Moreover, since the seed user of the second advertiser is usually the history seed user of the second advertiser, and the history seed user has been promoted the advertisement of the advertiser, the user information of the user who has successfully promoted the product of the second advertiser and has not performed the specified operation among the users who are the promotion objects of the product of the second advertiser can be obtained, and the second negative user information is obtained. The successful promotion specifically includes an advertisement about the product of the second advertiser, and the successful promotion is pushed or displayed to the user who is the promotion object of the product of the second advertiser.

Similar to the first positive user information and the first negative user information, since the specified operation is an operation on advertising conversion of the advertiser's product, the second positive user information may indicate information of users who are relatively more interested in the advertisement of the second advertiser, and the second negative user information may indicate information of users who are relatively not interested in the advertisement of the second advertiser. On the basis, the second positive user information and the second negative user information are used as the second user information of the seed user of the second advertiser, so that the subsequent expansion of users which are relatively easy to accept advertisements and are correspondingly more likely to be interested in the advertisements to be promoted from the user resource library is facilitated.

In an optional implementation manner, the crowd expansion based on the second user information to obtain fourth user information of a target user corresponding to the second user information may specifically include the following steps:

acquiring information about the advertisement conversion result of the product of the advertiser and information capable of influencing the advertisement conversion result of the product of the advertiser from second positive user information and second negative user information of the second user information;

constructing a second user characteristic of a seed user of a second advertiser using the obtained information;

inputting the pre-stored user information in the user resource library into a second model to obtain a second classification result of the pre-stored user information in the user resource library;

In a particular application, information regarding the advertisement conversion results of the advertiser's products, as well as information that can affect the advertisement conversion results of the advertiser's products, may be varied. For example, since the advertiser of the second user is different from the advertiser of the first user, in order to reduce omission of potential users, for the second user, when the product of the advertiser is APP, the information on the advertisement conversion result of the product of the advertiser may include: name of APP installed by the user. Also, the information that can affect the advertisement conversion result of the advertiser's product may include: information such as the brand and size of the user's electronic device, e.g., smart mobile terminal.

And, similar to the first user characteristic: and constructing the second user characteristics of the seed user of the second advertiser by using the obtained information, wherein the obtained information is spliced to obtain the second user characteristics, or the obtained information is combined into an array to obtain the second user characteristics, and the like. Similar to the first model: when a second model for obtaining a second classification result of the user is obtained by training using the second user characteristic, in order to deal with the characteristic that the data volume of the second user characteristic of the seed user of the second advertiser is large, the logistic regression model can be selected as the trained model to improve the training efficiency, and then the model parameter of the logistic regression model is adjusted through training to obtain the second model. In addition, in order to deal with the characteristic that the feature dimension of the second user feature is relatively high, the regular regression function can be added in an addition form in the cost function used in the logistic regression model training process, so that the problem that the second model obtained through training is over-fitted on a certain feature, and only user information with the feature is selected in a concentrated mode is solved, and the accuracy of crowd expansion is further improved.

As shown in fig. 2, a flow of a crowd expansion method for advertisement according to another embodiment of the present invention may include the following steps:

s201, first user information of a seed user of a first advertiser is obtained, and second user information of a seed user of a second advertiser is obtained.

S202, carrying out crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information, and carrying out crowd expansion based on the second user information to obtain fourth user information of the target user corresponding to the second user information.

The above-mentioned steps S201 to S202 are the same as the steps S101 to S102 in the embodiment of fig. 1 of the present invention, and are not repeated herein, for details, see the description of the embodiment of fig. 1 of the present invention.

And S203, promoting the advertisement to be promoted to the initial user indicated by the third user information and the initial user indicated by the fourth user information, and acquiring operation data of the advertisement to be promoted by each initial user.

Wherein the operation data comprises: data indicating whether the advertisement to be promoted is successfully promoted, and data indicating whether the advertisement to be promoted is subjected to specified operation.

S203 is similar to the alternative embodiment of the present invention related to step S103 in the embodiment of fig. 1, except that the operation data defined in S203 includes: data indicating whether the advertisement to be promoted is successfully promoted, and data indicating whether the advertisement to be promoted is subjected to specified operation. For the same parts, detailed description is omitted here, and the description of the alternative embodiment of fig. 1 of the present invention is given above. The data indicating whether the advertisement to be promoted is successfully promoted may include: and the electronic equipment of the user for receiving the advertisement to be promoted is used for receiving the data about whether the advertisement to be promoted is successfully displayed or pushed. The data indicating whether the advertisement to be promoted is subjected to the designated operation may include data indicating whether a product in the advertisement to be promoted is installed or activated or purchased.

And S204, when the operation data indicate that the number of the users who are to be promoted successfully promote the advertisement in each initial user is equal to the first number, obtaining the user information of the user who performs the specified operation in each initial user, and obtaining the third positive user information.

Wherein the first number may be the same as the number of users represented by the first positive user information. Illustratively, when a product in the advertisement to be promoted is an APP, user information of a user who installs or activates the APP in each initial user may be obtained when the advertisement to be promoted is successfully promoted to a first number of users, so as to obtain third positive user information.

S205, user information of users who are successful in advertisement promotion and do not perform specified operation in all initial users is obtained, and third negative user information is obtained.

Illustratively, when a product in the advertisement to be promoted is an APP, when the advertisement to be promoted is successfully promoted to a first number of users, user information of users who are not installed and activated in each initial user and who are only successfully promoted is obtained, and third negative user information is obtained.

And S206, respectively taking the users corresponding to the third positive user information and the third negative user information as seed users of the advertisement to be promoted.

And S207, performing crowd expansion based on the seed users of the advertisements to be promoted to obtain target users of the advertisements to be promoted.

S207 is the same as S104 in the embodiment of fig. 1, and is not repeated herein, for details, see the description of the embodiment of fig. 1.

In this optional embodiment, since the specified operation is an operation on advertisement conversion of the product of the advertiser, the third positive user information may indicate information of a user who is relatively more interested in the advertisement to be pushed, and the third negative user information may indicate information of a user who is relatively not interested in the advertisement to be pushed. On the basis, the users corresponding to the third positive user information and the third negative user information are used as seed users of the advertisements to be promoted, so that users which are relatively more likely to be interested in the advertisements to be promoted can be expanded from the user resource library, and the crowd expansion accuracy is improved.

In an optional implementation manner, the crowd expansion is performed on the seed users based on the advertisement to be promoted to obtain target users of the advertisement to be promoted, which may specifically include the following steps:

acquiring information about an advertisement conversion result of the advertisement to be promoted and information which can influence the advertisement conversion result of the advertisement to be promoted in third positive user information and third negative user information of seed users of the advertisement to be promoted;

training to obtain a third model for obtaining a third classification result of the user by utilizing the third user characteristic; wherein the third classification result comprises: a third positive target type having a user characteristic indicated by the third positive user information or a third negative target type having a user characteristic indicated by the third negative user information;

inputting the user information in the pre-stored user resource library into a third model to obtain a third classification result of the user information in the pre-stored user resource library;

Since the advertiser corresponding to the first user information is the same as the advertiser of the seed user of the advertisement to be promoted, the information about the advertisement conversion result of the advertisement to be promoted and the information that can affect the advertisement conversion result of the advertisement to be promoted, among the third positive user information and the third negative user information, are similar to the information for obtaining the first user characteristic. When a product in the advertisement to be promoted is APP, the information on the advertisement conversion result of the advertiser's product may include: the user installs the names of the APPs in a specified time period, for example, in the last three months, unloads the names of the APPs, retains the names of the APPs, and uses the names of the APPs with the frequency greater than a preset threshold. Also, the information that can affect the advertisement conversion result of the advertiser's product may include: the electronic equipment of the user comprises information such as the brand, the size, the name of an operator and the resident province of the user of the intelligent mobile terminal.

Or, for example, when the product in the advertisement to be promoted is an item sold by the e-commerce platform, the information on the advertisement conversion result of the product of the advertiser may include: the name of the item purchased by the user within a specified time period, such as the last three months, the name of the item marked as not interesting, the name of the collected item, and the name of the shared item. Also, the information that can affect the advertisement conversion result of the advertiser's product may include: the electronic device of the user is information such as the brand, size, operator name, resident province of the user, and the like of the smart mobile terminal.

And, similar to the first user characteristic, the third user characteristic of the seed user of the advertisement to be promoted is constructed by using the acquired information, which may be to splice the acquired information to obtain the third user characteristic, or to form an array of the acquired information to obtain the third user characteristic, and so on. Similar to the first model, when the third model for obtaining the third classification result of the user is obtained by training using the third user characteristic, in order to deal with the characteristic that the data volume of the third user characteristic of the seed user of the advertisement to be promoted is large, the logistic regression model can be selected as the trained model to improve the training efficiency, and then the model parameter of the logistic regression model is adjusted through training to obtain the third model. In addition, in order to deal with the characteristic that the feature dimension of the third user feature is relatively high, the regular regression function can be added in an addition form in the cost function used in the training process of the logistic regression model, the overfitting of the trained third classification model on a certain feature is reduced, the problem that only user information with the feature is selected in a concentrated mode occurs, and the accuracy of crowd expansion is further improved.

When the crowd of the advertisement to be promoted is expanded by specifically applying the optional embodiment, the obtained expansion effect comprises the following steps: the recall rate (recall) is about 84%, and the accuracy rate (precision) is about 84%. The method comprises the steps of obtaining a recall rate, wherein the recall rate is the number of users interested in the advertisement to be promoted in the target users obtained through expansion divided by the total number of users in a user resource library, and the accuracy rate is the number of users interested in the advertisement to be promoted in the target users obtained through expansion divided by the total number of users of the target users obtained through expansion. Therefore, the embodiment of the invention can give consideration to the richness and the accuracy of crowd expansion of advertisements.

Corresponding to the method embodiment, the embodiment of the invention also provides a crowd expansion device for the advertisement.

As shown in fig. 3, the structure of the crowd expansion device for advertising according to an embodiment of the present invention may include:

a user information obtaining module 301, configured to obtain first user information of a seed user of a first advertiser and obtain second user information of a seed user of a second advertiser; the first advertiser is an advertiser of the advertisement to be promoted, and the second advertiser is an advertiser different from the first advertiser;

an initial user obtaining module 302, configured to perform crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information, and perform crowd expansion based on the second user information to obtain fourth user information of the target user corresponding to the second user information;

a seed user obtaining module 303, configured to obtain a seed user of the advertisement to be promoted based on the third user information and the fourth user information;

and the target user expansion module 304 is configured to perform crowd expansion based on the seed users of the advertisement to be promoted to obtain target users of the advertisement to be promoted.

Optionally, the user information obtaining module 301 is specifically configured to:

Optionally, the initial user obtaining module 302 is specifically configured to:

Optionally, the seed user obtaining module 303 is specifically configured to:

the seed user obtaining module 303 is specifically configured to:

As shown in fig. 4, the crowd expansion apparatus for advertising according to another embodiment of the present invention may include:

a user information obtaining module 401, configured to obtain first user information of a seed user of a first advertiser and obtain second user information of a seed user of a second advertiser; the first advertiser is an advertiser of the advertisement to be promoted, and the second advertiser is an advertiser different from the first advertiser;

an initial user obtaining module 402, configured to perform crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information, and perform crowd expansion based on the second user information to obtain fourth user information of the target user corresponding to the second user information;

a seed user obtaining module 403, configured to promote the advertisement to be promoted to the initial user indicated by the third user information and the initial user indicated by the fourth user information, and obtain operation data of each initial user on the advertisement to be promoted; based on the operation data, selecting users meeting preset promotion effect conditions from all initial users as seed users of the advertisements to be promoted;

a target user expansion module 404, comprising:

a feature constructing submodule 4041, configured to obtain, from the third positive user information and the third negative user information of the seed user of the advertisement to be promoted, information about an advertisement conversion result of the advertisement to be promoted, and information that can affect the advertisement conversion result of the advertisement to be promoted; constructing a third user characteristic of the seed user of the advertisement to be promoted by using the acquired information;

the model training submodule 4042 is configured to train to obtain a third model for obtaining a third classification result of the user by using the third user characteristic; wherein the third classification result comprises: a third positive target type having a user characteristic indicated by the third positive user information or a third negative target type having a user characteristic indicated by the third negative user information;

the user expansion sub-module 4043 is configured to input the user information in the pre-stored user resource library into the third model, so as to obtain a third classification result of the user information in the pre-stored user resource library; and selecting the user corresponding to the user information of which the obtained third classification result is the third positive target type as the target user of the advertisement to be promoted.

Corresponding to the above embodiment, an embodiment of the present invention further provides an electronic device, as shown in fig. 5, where the electronic device may include:

the system comprises a processor 501, a communication interface 502, a memory 503 and a communication bus 504, wherein the processor 501, the communication interface 502 and the memory complete mutual communication through the communication bus 504 through the 503;

a memory 503 for storing a computer program;

the processor 501 is configured to implement the steps of the crowd expansion method for advertisement in any of the above embodiments when executing the computer program stored in the memory 503.

The Memory may include a RAM (Random Access Memory) or an NVM (Non-Volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.

The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component.

The computer-readable storage medium provided by an embodiment of the present invention is embodied in an electronic device, and a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the crowd expansion method for advertising in any of the above embodiments are implemented.

In yet another embodiment of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the method of crowd expansion of advertisements as described in any of the above embodiments.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in or transmitted from a computer-readable storage medium to another computer-readable storage medium, for example, from a website, computer, server, or data center, over a wired (e.g., coaxial cable, fiber optic, DSL (Digital Subscriber Line), or wireless (e.g., infrared, radio, microwave, etc.) network, to another website, computer, server, or data center, to any available medium that is accessible by a computer or that is a data storage device including one or more integrated servers, data centers, etc. the available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD (Digital versatile Disc, digital versatile disc)), or a semiconductor medium (e.g.: SSD (Solid state disk)), etc.

In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the device and electronic apparatus embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference may be made to some descriptions of the method embodiments for relevant points.

The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for crowd expansion of advertisements, the method comprising:

2. The method of claim 1, wherein obtaining the first user information of the seed user of the first advertiser comprises:

3. The method according to claim 2, wherein the crowd expansion based on the first user information to obtain third user information of a target user corresponding to the first user information comprises:

4. The method of claim 1, wherein obtaining second user information of a seed user of a second advertiser comprises:

5. The method according to claim 4, wherein the crowd expansion based on the second user information to obtain fourth user information of a target user corresponding to the second user information comprises:

6. The method according to any one of claims 1 to 5, wherein the obtaining of the seed user of the advertisement to be promoted based on the third user information and the fourth user information comprises:

7. The method of claim 6, wherein the operational data comprises: data indicating whether the advertisement to be promoted is successfully promoted or not and data indicating whether the specified operation is performed on the advertisement to be promoted or not;

8. An advertising crowd expansion apparatus, comprising:

a seed user obtaining module, configured to obtain a seed user of the advertisement to be promoted based on the third user information and the fourth user information;

9. An electronic device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the bus; the memory is used for storing a computer program; the processor, configured to execute the program stored in the memory, to implement the method steps according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that a computer program is stored in the storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of the claims 1-7.