WO2017101734A1

WO2017101734A1 - Method and device for recommending content item

Info

Publication number: WO2017101734A1
Application number: PCT/CN2016/109056
Authority: WO
Inventors: 何琪; 姚伶伶; 杨栋; 刘柄蔚
Original assignee: 腾讯科技（深圳）有限公司
Priority date: 2015-12-15
Filing date: 2016-12-08
Publication date: 2017-06-22
Also published as: CN105447724B; CN105447724A

Abstract

A method and device for recommending a content item. The method comprises: acquiring click history data of a user, where the click history data is used for indicating whether a user performed a click operation with respect to a content item in each past exposure opportunity (201); acquiring a click rate of the user on the basis of the click history data (202); calculating an effect parameter of an exposure opportunity with respect to different types of content items (203); and recommending a content item to the user on the basis of the effect parameter of the exposure opportunity with respect to the different types of content items (204).

Description

Content item recommendation method and device

The present application claims priority to Chinese Patent Application No. 201510932893.8, filed on Dec.

Technical field

The present invention relates to the field of network technologies, and in particular, to a content item recommendation method and apparatus.

Background technique

With the development of network technology, the network and every aspect of user life are inextricably linked. Based on this connection, advertisers can publish content items such as online advertisements through the network platform, thereby achieving the purpose of propagating themselves.

In order to meet different promotional needs, the existing network services can be divided into multiple different types, and each type can be charged in different ways. For example, for online advertising, it can be divided into CPD (Cost Per Day) advertising and CPC (Cost Per Click) advertising. Among them, CPD advertisements are sold on a daily basis, generally for brand advertisers. These advertisers pay more attention, exposure and coverage, etc.; while CPC advertisements are charged by effect, advertisers are more concerned about effects, such as The click rate of the ad and the CVR (Click Value Rate).

However, in the recommendation process of different types of content items, only the user traffic is randomly divided, that is, in the process of browsing the webpage, the user only randomly determines which type should be recommended on the current page. The content item leads to the inability to meet the actual needs of different types of content items, and the recommendation effect is poor.

Summary of the invention

In order to solve the problem of the prior art, an embodiment of the present invention provides a content item recommendation method and apparatus. The technical solution is as follows:

In one aspect, a content item recommendation method is provided, the method comprising:

Obtaining historical click data of the user, wherein the historical click data is used to indicate the user Whether the content item is clicked in each historical exposure opportunity;

Obtaining a click rate of the user according to the historical click data;

Calculating an influence parameter of the exposure opportunity on various types of content items according to the click rate;

A content item is recommended for the user based on the impact parameters of the exposure opportunity for various types of content items.

In another aspect, a content item recommendation device is provided, the device comprising:

An obtaining module, configured to acquire historical click data of the user, wherein the historical click data is used to indicate whether the user performs a click operation on the content item in each historical exposure opportunity;

a click rate obtaining module, configured to obtain a click rate of the user according to the historical click data;

a calculation module, configured to calculate an influence parameter of the exposure opportunity on various types of content items according to the click rate;

A recommendation module is configured to recommend content for the user according to an influence parameter of the exposure opportunity for various types of content items.

The beneficial effects brought by the technical solutions provided by the embodiments of the present invention are:

By analyzing whether the user performs a click operation in the historical exposure opportunity, whether the user is likely to click on the exposed content item in the exposure opportunity, thereby being able to determine the influence parameter of the exposure opportunity on different types of content items, and finally according to different The influence parameter of the type content item provides the user with the choice of which type of content item to display in the exposure opportunity, so that the display of the content item is no longer performed randomly, but can maximize the effect of the exposure opportunity, providing a more Reasonable exposure opportunity allocation method satisfies the actual needs of different types of content items and improves the recommendation effect.

DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the present invention. Other drawings may also be obtained from those of ordinary skill in the art in light of the inventive work.

1 is a schematic diagram of an implementation environment architecture provided by an embodiment of the present invention;

2 is a flowchart of a content item recommendation method according to an embodiment of the present invention;

3 is a flowchart of a content item recommendation method according to an embodiment of the present invention;

4 is a schematic structural diagram of a content item recommendation apparatus according to an embodiment of the present invention;

FIG. 5 is a block diagram of a content item recommending apparatus according to an exemplary embodiment.

detailed description

The embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

In order to facilitate an understanding of embodiments of the invention, some nouns are explained herein:

For online advertising, there are many different forms depending on its promotional focus, such as performance ads and brand ads.

A performance ad is an ad that focuses on sales or promotional performance. It can be an ad that promotes an activity or a certain type of product. This type of advertisement needs to be measured based on whether the user performed certain operational actions after seeing the advertisement. The effectiveness of its ads. These operational actions may include: clicking, sharing, purchasing, or other operational actions that are beneficial to the advertiser's production and sales behavior.

Brand advertising is an advertisement that emphasizes the brand image of the brand and increases the market share of the brand. It highlights an advertisement that identifies the location of the brand in the minds of consumers.

For the above two types of advertisements, different billing modes may be used. For example, the effect advertisements may adopt CPC, CPA (Cost PER Action) or CPS (Cost Per Sale). Among them, CPC is generally charged by click, a valid click behavior can charge the advertiser a certain fee; and CPA generally deducts the advertiser's certain fee after the user sees the advertisement has further expected behavior; and CPS is in accordance with The sales generated after seeing the advertisement are charged a certain fee for the advertiser. Brand advertising can take CPM (Cost Per Mille, billing by thousand exposure) and so on. CPM deducts a certain amount of money from advertisers based on the total number of impressions and CPM bids, such as ads for friends.

Exposure opportunities refer to any scenario in which an advertisement can be pushed. When an advertisement is successfully displayed to the user, the number of exposures is incremented by one.

FIG. 1 is a schematic diagram of an implementation environment architecture provided by an embodiment of the present invention. Referring to FIG. 1, the content item recommendation method can be applied to the implementation environment architecture. The implementation environment architecture may include a content item database, a history click database, a server, and the like. Wherein, the content item database is used to store content items, which are stored The content item may be multiple types of content items. Of course, the content item database may also store the delivery requirements of each content item, such as the delivery period, the target population, and the like. The history click database can be used to store the user's historical click data, which can be statistically stored by the server and stored in the history click database for subsequent analysis of the click behavior of any user, the historical click data can include the corresponding Whether the content item under an exposure opportunity is subjected to data such as clicking, and of course, may include, for example, a click time period, a clicked content item type, and the like. The server may refer to a network platform server, an information interaction server, and the like. The network platform server may refer to any platform that provides network services, for example, a multimedia sharing platform. The information exchange server may be a server for providing an information interaction service, such as a social application server or an instant communication server, which is not specifically limited in this embodiment of the present invention. The implementation environment architecture also includes a mobile terminal and a PC (Personal Computer). The mobile terminal can be a smart phone, a tablet computer, etc., and a client corresponding to the server can be installed on the mobile terminal or the PC, so that The user can use the service provided by the server through the client, and the user can browse information on the client, for example, browsing the webpage, browsing pictures or content items provided on the platform, information sent by other users, and the like. In the embodiment of the present invention, any scene that can be pushed by the content item can be regarded as an exposure opportunity. For example, if the user opens any webpage, the content item push position on the webpage can be regarded as an exposure opportunity; or, when the user opens the dynamic display page of the friend, the dynamic display page of the friend can be regarded as an exposure. The embodiment of the present invention does not specifically limit this.

It should be noted that the server in the embodiment of the present invention may be an independent physical device, or may be a cluster device composed of multiple physical devices. Of course, it may also be one or more functional modules on any device. The embodiments of the present invention are not limited thereto.

FIG. 2 is a flowchart of a content item recommendation method according to an embodiment of the present invention. Referring to FIG. 2, the method includes:

201. Obtain historical click data of the user, where the historical click data is used to indicate whether the user clicks on the content item in each historical exposure opportunity.

202. Obtain a click rate of the user according to the historical click data.

203. Calculate the influence of the exposure opportunity on various types of content items according to the click rate. parameter.

204. Recommend a content item for the user according to an influence parameter of the exposure opportunity for each type of content item.

The method provided by the embodiment of the present invention can determine whether the exposure opportunity for different types of content items by analyzing whether the user may click on the exposed content item in the exposure opportunity according to whether the user performs a click operation in the historical exposure opportunity. The influence parameter finally selects which type of content item to display in the exposure opportunity according to the influence parameter for different types of content items, so that the display of the content item is no longer performed randomly, but can enable the effect of the exposure opportunity Maximization provides a more reasonable way of assigning exposure opportunities, meeting the actual needs of different types of content items, and improving the recommendation effect.

Optionally, the content item type includes at least a first type of content item and a second type of content item. The higher the click rate, the greater the impact parameter of the exposure opportunity for the second type of content item.

Optionally, the method further includes: determining whether the user is a new user; if the user is a new user, recommending the first type of content item for the user.

Optionally, according to the influence parameter of the exposure opportunity for each type of content item, recommending the content item for the user includes:

Comparing the influence parameter and the first threshold of the exposure opportunity for various types of content items;

Determining to recommend a first type of content item for the user when the impact parameter of the exposure opportunity for each type of content item is less than a first threshold;

When the exposure parameter of the exposure opportunity for each type of content item is greater than or equal to the first threshold, it is determined to recommend the second type of content item for the user.

Optionally, different types of content items have different billing modes.

Optionally, the content item of the first type is a brand advertisement, and the content item of the second type is an effect advertisement.

All of the above optional technical solutions may be combined to form an optional embodiment of the present disclosure, and will not be further described herein.

FIG. 3 is a flowchart of a content item recommendation method according to an embodiment of the present invention. Referring to FIG. 3, the method includes:

301. Obtain historical click data of the user, and the historical click data is used to indicate whether the user clicks on the content item in each historical exposure opportunity.

For a user, in the process of using the service provided by the operator, multiple exposure opportunities can be generated, such as the user browsing the website, viewing the circle of friends, popping the window, and the like. Users may have different choices for these exposures. For example, some users who like to view advertisements may click on the content items provided in the exposure opportunity, while some users do not like this type of interruption, they may not Clicking on the content item provided in the exposure opportunity, in order to determine the user's click tendency, each time the content item is provided to the user, the user may be recorded whether the content item has been clicked under the exposure opportunity, for each A historical exposure opportunity, you can generate a data item to record what the user did in the historical exposure.

The historical click data may be time-sensitive, for example, only the click data of the user within a preset time period from the current time is acquired. Since the user's tendency may change, the more recent click data can be used as the basis for determining the user's tendency to improve the accuracy of the tendency determination and ensure that the content item is delivered more accurately.

302. Determine whether the user is a new user. If the user is a new user, recommend a content item corresponding to the first type for the user.

The content item corresponding to the first type is the brand advertisement, and the content item corresponding to the second type is the effect advertisement. For the effect advertisement, the user click rate is more emphasized, and for the brand advertisement, the new user is more paid attention to. Coverage and exposure rate, when the user is a new user, in order to improve the promotion effect of the brand advertisement, the first type of content item can be recommended for the user.

The new user may refer to a newly registered user, and may also refer to a user who has not been recommended by the brand advertisement. This embodiment of the present invention does not specifically limit this. Of course, the new and old users may not be distinguished, but the recommendation is performed according to the process of the subsequent step 303, which is not limited by the embodiment of the present invention.

Optionally, different types of content items have different billing modes.

303. Click the data according to the history to obtain the click rate of the user.

The click rate can be calculated based on the ratio of clicks to unclicked on each exposure in the historical click data. For example, for a user, in his historical click data, only 1 of the 10 exposures was clicked. Once, and the remaining 9 times are not clicked, the user's click rate can be 10%.

304. Calculate, according to the click rate, an influence parameter of the exposure opportunity on different types of content items. The content item type includes at least a first type and a second type, and the higher the click rate, the larger the influence parameter of the exposure opportunity on the second type content item.

The impact parameter refers to the impact of an exposure opportunity on the revenue of the content item, and the impact may refer to exposure impact, benefit impact, and the like.

When a user's click rate is high, it indicates that the user is more likely to click to view when being exposed to a content item (such as a network advertisement or popup information), and therefore, the user can be determined to be The influence parameter of the content item corresponding to the second type is larger. Since the click rate has a small impact parameter for some content items that are concerned about exposure, such as brand advertising, brand advertising is concerned with user coverage, then for brand advertising, the impact of the click rate on its impact parameters is not Big.

Therefore, in the embodiment of the present invention, the calculation of the influence parameter can be calculated according to ECPM (Effective CPM). The ECPM is mainly used to unify advertisements of various billing modes to observe the display efficiency of advertisements from the perspective of advertisement exposure price. For example, the formula for calculating the influence parameters of exposure opportunities for brand advertisements is as follows:

(cost/Ad impressions)×1000 Formula One

Where cost is used to represent the fee charged by the internal algorithm of the advertising system and should be charged from the advertiser;

Ad Impressions is the number of impressions that give the user a successful impression of an ad to produce an Ad impression.

In this embodiment of the invention, the influence parameter of the exposure opportunity on the effect advertisement can be calculated by the following formula:

N×cost’×1000 Formula 2

Where N represents the user's click-through rate for the content item, and cost' represents the fee paid by the advertiser for a single-click content.

305. Determine, according to an influence parameter of the exposure opportunity on different types of content items and a preset influence parameter balance point, determine a content item type that can satisfy the balance point of the preset influence parameter.

The preset influence parameter balance point may refer to an allocation percentage corresponding to different types of content items. For example, for performance ads and brand ads, you can assign 30% of the users to the performance ads, and 70% of the users to the brand ads to maximize the revenue of the platform, and through the calculation process of the above-mentioned influence parameters, you can determine the allocation to The 30% of users of performance ads are relatively high click-through users, so it’s important to analyze the user’s clickthrough rate, which will make conversions more likely. By assigning an effect advertisement to a user, the revenue generated by the brand advertisement for the advertisement operator can be improved without affecting the revenue generated by the brand advertisement, and the advertiser is improved in the expected effect. A win-win situation allows the advertising operator to use the limited network resources to get the maximum benefit, which is equivalent to improving the utilization of network resources.

The above steps 304 to 305 are processes for calculating the influence parameters of the exposure opportunity for different types of content items according to the history click data of the user. The advertising operator can determine the most suitable preset influence parameter balance point by measuring the balance points of different influence parameters. Of course, the preset influence parameter balance point can also be performed according to the actual distribution situation and the conversion rate during the distribution process. The embodiment of the present invention does not specifically limit this.

It should be noted that, according to the conversion rate of different users, a content item corresponding to the second type may be allocated to a user whose conversion rate is greater than a certain value, and a content item corresponding to the first type may be allocated to a user whose conversion rate is less than a certain value. To simplify the calculation process of the above influence parameters.

306. In the exposure opportunity, recommend a content item corresponding to the content item type for the user.

Applying this content item recommendation method to advertisements, by measuring the click conversion tendency of different users, the users who are easy to generate conversion are divided into effect advertisements, and the first-exposure user switches to the brand advertisement, which better satisfies the effect of the advertiser. Meet the needs of brand advertisers for crowd coverage while clicking on clicks and conversion needs.

FIG. 4 is a schematic structural diagram of a content item recommendation apparatus according to an embodiment of the present invention. Referring to Figure 4, the apparatus includes:

The obtaining module 401 is configured to acquire historical click data of the user, where the historical click data is used to indicate whether the user clicks on the content item in each historical exposure opportunity. Work

The click rate obtaining module 402 is configured to obtain the click rate of the user according to the historical click data;

The calculating module 403 is configured to calculate, according to the click rate, an influence parameter of the exposure opportunity on various types of content items;

The recommendation module 404 is configured to recommend a content item for the user according to the influence parameter of the exposure opportunity for various types of content items.

Optionally, each type of content item includes at least a first type of content item and a second type of content item, wherein an impact parameter of the exposure opportunity for the second type of content item follows the click rate Increase and increase.

Optionally, the device further includes: a determining module, configured to determine whether the user is a new user; and the recommending module is further configured to: if the determining module determines that the user is a new user, recommend the user A type of content item.

Optionally, the recommendation module is configured to compare an impact parameter of the exposure opportunity with various types of content items and a first threshold; when the exposure parameter of the exposure opportunity for each type of content item is less than a first threshold, determining Determining a first type of content item for the user; and determining to recommend a second type of content item for the user when the exposure parameter has an impact parameter for each type of content item that is greater than or equal to a first threshold.

Optionally, different types of content items have different billing modes.

It should be noted that the content item recommendation device provided by the foregoing embodiment is only illustrated by the division of the foregoing functional modules when the content item is recommended. In actual applications, the function distribution may be completed by different functional modules as needed. The internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the content item recommendation device and the content item recommendation method embodiment provided by the foregoing embodiments are in the same concept, and the specific implementation process is described in detail in the method embodiment, and details are not described herein again.

FIG. 5 is a block diagram of a content item recommendation device 500, according to an exemplary embodiment. example For example, device 500 can be provided as a server. Referring to FIG. 5, apparatus 500 includes a processing component 522 that further includes one or more processors, and memory resources represented by memory 532 for storing instructions executable by processing component 522, such as an application. An application stored in memory 532 can include one or more modules each corresponding to a set of instructions. Further, processing component 522 is configured to execute instructions to perform the content item recommendation method described above.

Apparatus 500 can also include a power supply component 526 configured to perform power management of apparatus 500, a wired or wireless network interface 550 configured to connect apparatus 500 to the network, and an input/output (I/O) interface 558. Device 500 can operate based on an operating system stored in the memory 532, such as ^{^{Windows Server TM, Mac OS X TM}} , Unix TM, Linux TM, FreeBSD TM or the like.

A person skilled in the art may understand that all or part of the steps of implementing the above embodiments may be completed by hardware, or may be instructed by a program to execute related hardware, and the program may be stored in a computer readable storage medium. The storage medium mentioned may be a read only memory, a magnetic disk or an optical disk or the like.

The above are only the preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalents, improvements, etc., which are within the spirit and scope of the present invention, should be included in the protection of the present invention. Within the scope.

Claims

A content item recommendation method, the method comprising:

Obtaining historical click data of the user, wherein the historical click data is used to indicate whether the user performs a click operation on the content item in each historical exposure opportunity;

Obtaining a click rate of the user according to the historical click data;

Calculating an influence parameter of the exposure opportunity for various types of content items according to the click rate; and recommending the content item for the user according to the influence parameter of the exposure opportunity for various types of content items.
The method of claim 1 wherein said various types of content items comprise at least a first type of content item and a second type of content item, wherein said exposure opportunity is for said second type of content item The influence parameter of the content item increases as the click rate increases.
The method of claim 2, further comprising:

Determining whether the user is a new user;

If the user is a new user, recommend the first type of content item for the user.
The method according to claim 2, wherein the recommending the content item for the user according to the influence parameter of the exposure opportunity on the various types of content items comprises:

Comparing the influence parameter and the first threshold of the exposure opportunity for various types of content items;

Determining to recommend a first type of content item for the user when the impact parameter of the exposure opportunity for each type of content item is less than a first threshold;

When the exposure parameter of the exposure opportunity for each type of content item is greater than or equal to the first threshold, it is determined to recommend the second type of content item for the user.
The method according to any of the preceding claims, characterized in that the different types of content items have different charging modes.
The method of claim 2 wherein the first type of content item is a brand advertisement and the second type of content item is an effect advertisement.
A content item recommendation device, wherein the device comprises:

An obtaining module, configured to acquire historical click data of the user, wherein the historical click data is used to indicate whether the user performs a click operation on the content item in each historical exposure opportunity;

a click rate obtaining module, configured to obtain a click rate of the user according to the historical click data;

a calculation module, configured to calculate an influence parameter of the exposure opportunity on various types of content items according to the click rate;

A recommendation module is configured to recommend content for the user according to an influence parameter of the exposure opportunity for various types of content items.
The apparatus of claim 7, wherein the various types of content items comprise at least a first type of content item and a second type of content item, wherein the exposure opportunity is for the second type of content item The influence parameter of the content item increases as the click rate increases.
The device according to claim 8, wherein the device further comprises:

a determining module, configured to determine whether the user is a new user;

The recommendation module is further configured to: if the determining module determines that the user is a new user, recommend a first type of content item for the user.
The apparatus according to claim 8, wherein said recommendation module is further configured to compare an influence parameter of said exposure opportunity with respect to various types of content items and a first threshold; when said exposure opportunity is for each type of content item When the impact parameter is less than the first threshold, determining to recommend the first type of content item for the user; and determining that the user recommendation is when the impact parameter of the exposure opportunity for the various types of content items is greater than or equal to the first threshold The second type of content item.
Device according to any of claims 7-10, characterized in that within different types The tolerances have different billing modes.
The apparatus of claim 8, wherein the first type of content item is a brand advertisement and the second type of content item is an effect advertisement.
A non-volatile storage medium for storing one or more computer programs, wherein the computer program includes processor-executable instructions having one or more memories; the processor runs the instructions to perform The content item recommendation method according to any one of claims 1 to 6.