Advertisement recommendation method, system, electronic equipment and device

Abstract

The invention provides an advertisement recommendation method, system, electronic equipment and device, and the method comprises the steps: receiving an advertisement request sent by second equipment, and determining a target value type corresponding to the advertisement request based on the search content in the advertisement request; performing advertisement recommendation on the advertisement request based on a target value type corresponding to the advertisement request to obtain N search advertisements corresponding to the target value type, N being a positive integer; and sending the N search advertisements to the second equipment. According to the method provided by the invention, the recommendation effect of the advertisement platform on the searched advertisement is improved, so that the recommended advertisement better conforms to the intention of the user, and the satisfaction degree of the user can be improved.

Description

Technical Field

[0001] This application relates to the field of information technology, and in particular, to an advertisement recommendation method, system, electronic device, and apparatus. Background Art

[0002] Search Advertisements (Search Ads) refer to an advertisement link method in which, after an advertiser places advertisement materials and keywords through an advertisement platform, a user can query by actively entering a search term and relevant advertisements are returned.

[0003] It can be understood that search advertisements are advertisements obtained by a user through advertisement search on an advertisement platform, and display advertisements are advertisements actively displayed by the advertisement platform. When the search terms of a user are different, the strength of the user's intention for advertisements is also different. In addition, compared with display advertisements, search advertisements have a higher correlation with the user's intention. Therefore, search advertisements have higher value than display advertisements, and their value can be reflected in a higher average cost per mille (CPM).

[0004] However, in the current advertisement platform during the recommendation process of search advertisements, since the advertisements recommended by the advertisement platform have a low matching degree with the user's intention and the recommendation effect is poor, it is difficult for the advertisement platform to identify high-value advertisements, and thus it is impossible to set a reasonable bidding strategy for high-value advertisements, resulting in insufficient advertisement bidding, causing losses to the advertisement platform and reducing the satisfaction of advertisers. Summary of the Invention

[0005] This application provides an advertisement recommendation method, system, electronic device, and apparatus, which helps to improve the recommendation effect of search advertisements on the advertisement platform, makes the recommended advertisements more in line with the user's intention, and thus can improve user satisfaction.

[0006] In a first aspect, this application provides an advertisement recommendation method, which is applied to a first device and includes: receiving an advertisement request sent by a second device, where the advertisement request includes search content; determining a target value type corresponding to the advertisement request based on the search content in the advertisement request; performing advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request to obtain N search advertisements corresponding to the target value type, where N is a positive integer; and sending the N search advertisements to the second device.

[0007] In this application, after receiving an advertisement request, the first device identifies the value type of the advertisement request, and thus can perform advertisement recommendation on the advertisement request corresponding to the identified value type in a targeted manner, making the recommended advertisements more in line with the user's intention to improve user satisfaction.

[0008] In one possible implementation, the method further includes: using the result of determining the target value type corresponding to the advertisement request as a virtual advertisement slot.

[0009] In this application, the first device sets different virtual advertisement slots for advertisements recommended for advertisement requests of different value types, and can display them for advertisers to view, so that advertisers can intuitively and specifically understand the value of the advertisement request. In addition, the advertisement platform can set different bidding strategies for different virtual advertisement slots, enabling advertisers to bid based on different virtual advertisement slots to increase the revenue of the advertisement platform.

[0010] In one possible implementation, the step of performing advertisement recommendation for the advertisement request based on the target value type corresponding to the advertisement request to obtain N search advertisements corresponding to the target value type includes: determining the target advertisement type corresponding to the advertisement request based on the search content in the advertisement request; performing advertisement recommendation for the advertisement request based on the target value type and the target advertisement type corresponding to the advertisement request to obtain N search advertisements corresponding to the target value type and the target advertisement type.

[0011] In this application, the first device can recommend advertisements corresponding to the advertisement type by identifying the advertisement type required by the user, making the recommended advertisements more in line with the user's intention and improving the user's satisfaction.

[0012] In one possible implementation, the step of determining the target advertisement type corresponding to the advertisement request based on the search content in the advertisement request includes: determining the user intention based on the search content in the advertisement request; determining the target advertisement type corresponding to the advertisement request based on the user intention.

[0013] In one possible implementation, the advertisement type at least includes application advertisements and commodity advertisements.

[0014] In one possible implementation, the N search advertisements are the top N ranked search advertisements, and the ranking of the N search advertisements is determined by the estimated earnings per thousand impressions ECPM.

[0015] In one possible implementation, the N search advertisements are obtained after ranking based on the first candidate search advertisement and the second candidate search advertisement. Among them, the first candidate search advertisement is obtained after performing advertisement recommendation in the first device, and the second candidate search advertisement is obtained after performing advertisement recommendation in the third device, and the third device is the device of the third-party demand-side platform DSP.

[0016] In this application, the first device comprehensively considers the rankings of the recommended advertisements of its own DSP and the recommended advertisements of the third-party DSP, so that the top N search advertisements finally output can take into account the rankings of multiple DSPs and reduce the errors caused by the recommendations of a single DSP.

[0017] In one possible implementation, different value types correspond to different recommendation models; the advertising recommendation for the advertising request based on the target value type corresponding to the advertising request to obtain N search advertisements corresponding to the target value type includes: determining a target recommendation model corresponding to the target value type based on the target value type corresponding to the advertising request; using the target recommendation model corresponding to the target value type to perform advertising recommendation on the advertising request to obtain N search advertisements corresponding to the target value type.

[0018] In this application, the first device uses different recommendation models to perform advertising recommendations for advertising requests of different value types, so that advertising requests of different value types can be processed quickly and accurately to obtain the required recommended advertisements.

[0019] In one possible implementation, the determining the target value type corresponding to the advertising request based on the search content in the advertising request includes: identifying the search content in the advertising request to obtain keywords; performing text matching between the keywords and the list of advertisements in progress to determine the target value type corresponding to the advertising request.

[0020] In this application, the first device can quickly determine the value type corresponding to the advertising request by performing text matching between the keywords in the search content and the list of advertisements in progress.

[0021] In one possible implementation, the method further includes: periodically updating the list of advertisements in progress.

[0022] In this application, the first device can avoid the errors caused by the changes in the list of advertisements in progress by periodically updating the list of advertisements in progress.

[0023] In one possible implementation, the text matching includes multiple results; the multiple results have a one-to-one mapping relationship with multiple value types; or, the multiple results have a many-to-one mapping relationship with multiple value types.

[0024] Second aspect, the present application provides an advertisement recommendation method, which is applied to an advertisement recommendation system. The advertisement recommendation system includes a first device and a second device. The method includes: the second device sends an advertisement request to the first device in response to search content input by a user; the first device receives the advertisement request sent by the second device, where the advertisement request includes the search content; the first device determines a target value type corresponding to the advertisement request based on the search content in the advertisement request; the first device performs advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request, and obtains N search advertisements corresponding to the target value type, where N is a positive integer; the first device sends the N search advertisements to the second device; the second device displays the N search advertisements.

[0025] In the present application, after the second device sends an advertisement request to the first device in response to the search content input by the user, the first device receives the advertisement request and identifies the value type of the advertisement request. Thus, it is possible to perform advertisement recommendation on the advertisement request corresponding to the identified value type in a targeted manner, making the recommended advertisements more in line with the user's intention and presenting them to the user on the second device to improve user satisfaction.

[0026] In one possible implementation, the method further includes: the first device uses the result of determining the target value type corresponding to the advertisement request as a virtual advertisement slot.

[0027] In the present application, the first device sets different virtual advertisement slots for the advertisements recommended for advertisement requests of different value types and can display them for advertisers to view, so that advertisers can intuitively and specifically understand the value of the advertisement request. In addition, the advertisement platform can set different bidding strategies for different virtual advertisement slots, enabling advertisers to bid based on different virtual advertisement slots to increase the revenue of the advertisement platform.

[0028] In one possible implementation, the first device performs advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request, and obtains N search advertisements corresponding to the target value type, including: the first device determines a target advertisement type corresponding to the advertisement request based on the search content in the advertisement request; the first device performs advertisement recommendation on the advertisement request based on the target value type and the target advertisement type corresponding to the advertisement request, and obtains the N search advertisements corresponding to the target value type and the target advertisement type.

[0029] In this application, the first device can recommend advertisements corresponding to the advertisement type by identifying the advertisement type required by the user, making the recommended advertisements more in line with the user's intention and improving the user's satisfaction.

[0030] In one possible implementation, the first device determines the target advertisement type corresponding to the advertisement request based on the search content in the advertisement request, including: the first device determines the user intention based on the search content in the advertisement request; the first device determines the target advertisement type corresponding to the advertisement request based on the user intention.

[0031] In this application, the first device can recommend advertisements corresponding to the advertisement type by identifying the advertisement type required by the user, making the recommended advertisements more in line with the user's intention and improving the user's satisfaction.

[0032] In one possible implementation, the advertisement type at least includes application advertisements and commodity advertisements.

[0033] In one possible implementation, the N search advertisements are the top N search advertisements, and the ranking of the N search advertisements is determined by the estimated earnings per thousand impressions (ECPM).

[0034] In one possible implementation, the N search advertisements are obtained after ranking based on the first candidate search advertisement and the second candidate search advertisement. Among them, the first candidate search advertisement is obtained after advertisement recommendation in the first device, and the second candidate search advertisement is obtained after advertisement recommendation in the third device. The third device is the device of the third-party demand-side platform DSP.

[0035] In this application, by comprehensively considering the ranking of the recommended advertisements of its own DSP and the ranking of the recommended advertisements of the third-party DSP, the first device enables the finally output top N search advertisements to take into account the rankings of multiple DSPs and reduce the errors caused by the recommendation of a single DSP.

[0036] In one possible implementation, different value types correspond to different recommendation models; for the advertisement recommendation of the advertisement request based on the target value type corresponding to the advertisement request to obtain N search advertisements corresponding to the target value type, it includes: determining the target recommendation model corresponding to the target value type based on the target value type corresponding to the advertisement request; using the target recommendation model corresponding to the target value type to perform advertisement recommendation on the advertisement request to obtain N search advertisements corresponding to the target value type.

[0037] In this application, the first device uses different recommendation models to perform advertisement recommendations for advertisement requests of different value types, so that advertisement requests of different value types can be processed quickly and accurately to obtain the required recommended advertisements.

[0038] In one possible implementation, determining the target value type corresponding to the advertisement request based on the search content in the advertisement request includes: identifying the search content in the advertisement request to obtain keywords; performing text matching between the keywords and the list of advertisements being served to determine the target value type corresponding to the advertisement request.

[0039] In this application, the first device can quickly determine the value type corresponding to an advertisement request by performing text matching between the keywords in the search content and the list of advertisements being served.

[0040] In one possible implementation, the method further includes: periodically updating the list of advertisements being served.

[0041] In this application, the first device can avoid errors caused by changes in the list of advertisements being served by periodically updating the list of advertisements being served.

[0042] In one possible implementation, the text matching includes multiple results; the multiple results have a one-to-one mapping relationship with multiple value types; or, the multiple results have a many-to-one mapping relationship with multiple value types.

[0043] In a third aspect, this application provides an advertisement recommendation system, including a first device and a second device. Among them, the first device implements the advertisement recommendation method executed by the first device as described in the second aspect, or, the first device implements the advertisement recommendation method as described in the first aspect; the second device implements the advertisement recommendation method executed by the second device as described in the second aspect.

[0044] In a fourth aspect, this application provides an electronic device, including: a memory, one or more processors, and one or more programs; wherein the one or more programs are stored in the memory, and when the one or more processors execute the one or more programs, the electronic device implements the advertisement recommendation method as described in the first aspect.

[0045] In a fifth aspect, this application provides an advertisement recommendation device, including: one or more functional modules, and the one or more functional modules are used to execute any one of the advertisement recommendation methods provided in the first aspect.

[0046] Sixth aspect, the present application provides a chip system, which includes a processing circuit and a storage medium, and computer program code is stored in the storage medium; when the computer program code is executed by the processing circuit, the advertising recommendation method described in the first aspect or the second aspect is implemented.

[0047] Seventh aspect, the present application provides a readable storage medium, in which a program is stored, and when it runs on a first device, the first device is enabled to implement the advertising recommendation method described in the first aspect or the second aspect.

[0048] Eighth aspect, the present application provides a program, and when the above program runs on the processor of a first device, the first device is enabled to execute the advertising recommendation method described in the first aspect.

[0049] In a possible design, the program in the eighth aspect can be stored in whole or in part on a storage medium packaged together with the processor, or can be stored in whole or in part on a memory not packaged together with the processor. Description of the Drawings

[0050] Figure 1 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application;

[0051] Figure 2 It is a schematic system architecture diagram provided by an embodiment of the present application;

[0052] Figure 3 It is a schematic flowchart of an embodiment of the advertising recommendation method provided by the present application;

[0053] Figure 4A and Figure 4B It is a schematic diagram of the display of a search advertisement provided by an embodiment of the present application;

[0054] Figure 5 It is a schematic diagram of the value type of an advertisement request provided by an embodiment of the present application;

[0055] Figure 6 It is a schematic diagram of a search advertisement returned by a first device provided by an embodiment of the present application;

[0056] Figure 7 It is a schematic diagram of a second device displaying a search advertisement provided by an embodiment of the present application;

[0057] Figure 8 It is a schematic system architecture diagram on the server side provided by an embodiment of the present application;

[0058] Figure 9 It is a schematic flowchart of another embodiment of the advertising recommendation method provided by the present application;

[0059] Figure 10 Schematic structural diagram of an embodiment of the advertisement recommendation device provided for this application. Detailed implementation manners

[0060] In the embodiments of this application, unless otherwise specified, the character " / " indicates that the associated objects before and after are in an "or" relationship. For example, A / B may represent A or B. "And / or" describes the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone, these three situations.

[0061] It should be noted that the terms "first", "second", etc. involved in the embodiments of this application are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features, nor can they be understood as indicating or implying an order.

[0062] In the embodiments of this application, "at least one" means one or more, and "a plurality" means two or more. In addition, "at least one (item)" or its similar expressions refer to any combination of these items, which can include any combination of single item (item) or plural items (items). For example, at least one (item) of A, B, or C can represent: A, B, C, A and B, A and C, B and C, or A, B, and C. Among them, each of A, B, and C can itself be an element or a set containing one or more elements.

[0063] In the embodiments of this application, "exemplary", "in some embodiments", "in another embodiment", etc. are used to give examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary" in this application should not be interpreted as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of the word "exemplary" aims to present concepts in a specific way.

[0064] In the embodiments of this application, "(of)", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when not emphasizing their differences, the meanings to be expressed are the same. In the embodiments of this application, communication and transmission can sometimes be used interchangeably. It should be noted that when not emphasizing their differences, the meanings they express are the same. For example, transmission can include sending and / or receiving, and can be a noun or a verb.

[0065] In the embodiments of this application, the equality involved can be used in combination with greater than, applicable to the technical solutions adopted when it is greater than, or can also be used in combination with less than, applicable to the technical solutions adopted when it is less than. It should be noted that when equality is used in combination with greater than, it cannot be used in combination with less than; when equality is used in combination with less than, it is not used in combination with greater than.

[0066] Search Advertisements (Search Ads) refers to an advertising link method in which, after an advertiser places advertising materials and keywords through an advertising platform, users can query by actively entering search terms and relevant advertisements will be returned.

[0067] It can be understood that search ads are ads obtained by users through search in the advertising platform in the way of advertising search, and display ads are ads actively displayed by the advertising platform. When the search terms of users are different, the strength of users' intentions for ads is also different. In addition, compared with display ads, search ads have a higher correlation with users' intentions. Therefore, search ads have higher value than display ads, and their value can be reflected in a higher average cost per mille (CPM).

[0068] However, in the current advertising platform during the recommendation process of search ads, due to the low matching degree between the ads recommended by the advertising platform and users' intentions, the recommendation effect is poor. As a result, it is difficult for the advertising platform to identify high-value ads, and thus it is impossible to set a reasonable bidding strategy for high-value ads, leading to insufficient ad bidding, causing losses to the advertising platform and reducing the satisfaction of advertisers.

[0069] Based on the above problems, the embodiments of the present application propose an advertising recommendation method applied to an electronic device. The electronic device can be a server in the advertising platform. Among them, the server can be a Windows server, a Linux server, or other server devices that can provide simultaneous access for multiple devices. The embodiments of the present application do not make special limitations on the type of the server.

[0070] In some alternative embodiments, the electronic device can also be a server cluster in the advertising platform. The server cluster can be a device cluster composed of multiple regions, multiple computer rooms, and multiple servers. The embodiments of the present application do not make special limitations on this.

[0071] It can be understood that the electronic device in the embodiments of the present application can support functions such as message storage and distribution, large-scale data storage, large-scale data processing, multi-concurrency and high-concurrency processing, and data redundancy backup.

[0072] Figure 1 First, a schematic structural diagram of the electronic device 100 is exemplarily shown.

[0073] The above-mentioned electronic device 100 may include: at least one processor; and at least one memory communicatively connected to the above-mentioned processor, where: the above-mentioned memory stores program instructions executable by the above-mentioned processor, and the processor can execute the methods provided by the embodiments shown in this article by calling the above-mentioned program instructions.

[0074] Figure 1 A block diagram showing an exemplary electronic device 100 suitable for implementing the embodiments herein is shown. Figure 1 The shown electronic device 100 is merely an example and should not impose any limitation on the functions and usage scope of the embodiments herein.

[0075] As Figure 1 shown, the components of the electronic device 100 may include, but are not limited to: one or more processors 110, a memory 120, a communication bus 140 connecting different system components (including the memory 120 and the processor 110), and a communication interface 130.

[0076] The communication bus 140 represents one or more of several types of bus structures, including a memory bus or a memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus structures. By way of example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnection (PCI) bus.

[0077] The electronic device 100 typically includes a variety of computer system readable media. These media can be any available media accessible by the device, including volatile and non-volatile media, removable and non-removable media.

[0078] The memory 120 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The device may further include other removable / non-removable, volatile / non-volatile computer system storage media. Although Figure 1Not shown in the figure, a disk drive for reading and writing a removable non-volatile disk (such as a "floppy disk") and an optical disk drive for reading and writing a removable non-volatile optical disk (such as a Compact Disc Read Only Memory (hereinafter referred to as CD-ROM), a Digital Video Disc Read Only Memory (hereinafter referred to as DVD-ROM), or other optical media) can be provided. In these cases, each drive can be connected to the communication bus 140 through one or more data medium interfaces. The memory 120 can include at least one program product, and the program product has a set (such as at least one) of program modules, and these program modules are configured to execute the functions of the embodiments herein.

[0079] A program / utility with a set (at least one) of program modules can be stored in the memory 120. Such program modules include - but are not limited to - an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment. The program modules generally execute the functions and / or methods in the embodiments described herein.

[0080] The electronic device 100 can also communicate with one or more external devices (such as a keyboard, a pointing device, a display, etc.), and can also communicate with one or more devices that enable a user to interact with the device, and / or communicate with any device that enables the device to communicate with one or more other devices (such as a network card, a modem, etc.). Such communication can be carried out through the communication interface 130. And, the electronic device 100 can also communicate with one or more networks (such as a Local Area Network (hereinafter referred to as LAN), a Wide Area Network (hereinafter referred to as WAN), and / or a public network, such as the Internet) through a network adapter ( Figure 1 not shown in the figure), and the above network adapter can communicate with other modules of the device through the communication bus 140. It should be understood that although Figure 1 not shown in the figure, other hardware and / or software modules can be used in combination with the electronic device 100, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, Redundant Arrays of Independent Drives (hereinafter referred to as RAID) systems, tape drives, and data backup storage systems, etc.

[0081] The processor 110 executes various functional applications and data processing by running the programs stored in the memory 120, such as implementing the methods provided by the embodiments herein.

[0082] It can be understood that the interface connection relationships among the modules illustrated in the embodiments of this article are only illustrative descriptions and do not constitute a structural limitation on the electronic device 100. In other embodiments of this article, the electronic device 100 may also adopt different interface connection methods in the above embodiments, or a combination of multiple interface connection methods.

[0083] Figure 2 This is the system architecture diagram provided by the embodiments of this application. As Figure 2 shown, the system architecture may include a first device 20 and a second device 21.

[0084] Among them, the first device 20 may be the above-mentioned electronic device 100. The first device 20 may receive an advertisement request from the second device 21 and may return a recommended search advertisement to the second device 21 according to the advertisement request.

[0085] It can be understood that in the embodiments of this application, after receiving the advertisement request from the second device, the first device 20 may determine the corresponding value based on the search content in the advertisement request, and may recommend an advertisement corresponding to the determined value according to the value, and return the sorted search advertisement to the second device 21.

[0086] The second device 21 may be a device on the terminal side (user side). Exemplarily, the second device 21 may include, but is not limited to, devices with a display screen such as a personal computer (PC), a mobile phone, a Pad, a tablet, etc. The embodiments of this application do not make special limitations on the type of the second device 21.

[0087] It can be understood that a target application may be installed on the second device 21. The target application may be used for advertisement placement and may be regarded as a medium between the user and the advertisement platform. The display interface of the target application may include a search box and a search advertisement display area. Among them, the search box may be used to input the user's search content, and the search advertisement display area may be used to display the search advertisement returned by the first device. For example, the second device 21 may send an advertisement request to the first device 20 through the target application, receive the search advertisement returned by the first device 20, and display the search advertisement returned by the first device 20 on the display interface of the second device 21.

[0088] In some alternative embodiments, in addition to the above-mentioned target application, other applications may also be installed on the second device 21. For example, camera, image management, message processing, data processing and sending, text processing, instant message push, network communication, media playback, and time management, etc.

[0089] It should be noted that the application scenarios of the embodiments of the present application may include, but are not limited to, scenarios such as search browsers or global searches.

[0090] Table 1 exemplarily shows the above application scenarios and their scenario descriptions.

[0091] Table 1

[0092]

[0093] The advertising delivery targets targeted by the embodiments of the present application may include, but are not limited to, wake-up, secondary retention, payment, activation, registration, form, or other custom user behaviors.

[0094] Table 2 exemplarily shows the specific information of the above delivery targets.

[0095] Table 2

[0096]

[0097] Next, in combination with Figure 3 , Figure 4A , Figure 4B , Figures 5 - 7 an exemplary description of the advertising recommendation method provided by the embodiments of the present application is given.

[0098] As Figure 3 shown is a schematic flowchart of an embodiment of the advertising recommendation method provided by the present application, which specifically includes the following steps:

[0099] Step 301, in response to the search content input by the user, the second device sends an advertisement request to the first device, where the advertisement request includes the search content. Correspondingly, the first device receives the advertisement request sent by the second device.

[0100] Specifically, when the user has a need for search advertising, the user can open the target application in the second device and can enter the search content in the search box on the display page of the target application.

[0101] It can be understood that the display page of the target application may include a search box, and the user can enter the search content in the search box. The specific manifestation form of the search box on the display page of the target application can refer to the display interface of the second device in Figure 2 , which will not be elaborated here.

[0102] Among them, the advertisement types of search advertising may include, but are not limited to, application advertisements and commodity advertisements. Application advertisements refer to the user requesting the server to return relevant applications through the search content. Commodity advertisements refer to the user requesting the server to return relevant commodities through the search content.

[0103] It can be understood that the above types of search ads are only illustrative and do not constitute a limitation on the embodiments of the present application. In some embodiments, search ads may also include other types, which will not be enumerated one by one here.

[0104] In response to the search ad input by the user, the second device may send an ad request to the first device, where the ad request may include the search content.

[0105] Exemplarily, referring to Figure 4A As shown, taking the search ad as an app ad as an example, the user may input the following search content: "Housing near the South Third Ring Road?" From this search content, it can be known that the user's intention may be to rent or buy a house. Therefore, this search content can be used to request the return of apps related to real estate.

[0106] Or, referring to Figure 4B As shown, taking the search ad as a product ad as an example, the user may input the following search content: "Popular summer outfits?" From this search content, it can be known that the user's intention may be to look for products related to fashion or outfits. Therefore, this search content can be used to request the return of products related to fashion or outfits, such as bags, clothes, accessories, shoes, etc.

[0107] In some optional embodiments, the ad type of the search ad may be indicated by the ad request. Exemplarily, the search box may correspond to the ad type of the search ad. After the user inputs the search content in any one of the search boxes, the second device sends an ad request to the first device. The ad request carrying the search content may also carry an ad type indication, which can be used to indicate to the first device the ad type corresponding to the search box where the user currently inputs the search content. Taking the ad types of the search ad including app ads and product ads as an example, the display page of the target app may include 2 search boxes, which may be an app ad search box and a product ad search box respectively. Among them, the app ad search box can be used to request the return of search ads of the app ad type, and the product ad search box can be used to request the return of search ads of the product ad type. When the user inputs the search content in the app ad search box and the second device sends an ad request to the first device, the ad request may include an ad type indication, which can indicate to the first device that the ad type currently requested by the user is the app ad type; or, when the user inputs the search content in the product ad search box and the second device sends an ad request to the first device, the ad request may include an ad type indication, which can indicate to the first device that the ad type currently requested by the user is the product ad type.

[0108] In some alternative embodiments, the ad type of a search ad can be determined by identifying the search content in an ad request. Exemplarily, the display page of a target application may include one search box, which can be used to request search ads of different ad types. For example, taking the ad types of search ads including app ads and product ads as an example, when a user enters search content in the search box and the second device sends an ad request to the first device, the first device can identify the search content in the ad request to determine the ad type of the search ad corresponding to the ad request, that is, determine whether the search for this ad request is for an app ad or a product ad.

[0109] Step 302, in response to the received ad request, the first device determines the target value type corresponding to the ad request based on the search content in the ad request.

[0110] Specifically, when the first device receives an ad request, the first device can identify the search content in the ad request to determine the target value type corresponding to the ad request. Among them, the value type can be used to characterize the high or low value of the current ad request.

[0111] Exemplarily, the value type may include but is not limited to low value and high value. The low value type can be used to characterize that the value of the current ad request is low, and the high value type can be used to characterize that the value of the current ad request is high.

[0112] It can be understood that the number of types of the above value types is only an exemplary illustration and does not constitute a limitation on the embodiments of the present application. In some embodiments, the number of types of the value type can be set according to actual needs. For example, the value type may include low value, medium value, high value, etc.

[0113] Now, taking the value type including low value and high value as an example and combining Figure 5 for exemplary illustration. Refer to Figure 5 , the multiple ad requests in the left column contain clear brand words. For example, and other brands. These brands can better express the ad intent required by the user. Therefore, the multiple ad requests in the left column can be considered high-value ad requests. The content in the multiple ad requests in the right column cannot accurately express the ad intent required by the user. Therefore, the multiple ad requests in the right column can be considered low-value ad requests.

[0114] In some alternative embodiments, the manner in which the first device identifies the search content in the advertisement request to determine the value type in the advertisement request may include: The first device identifies the search content in the advertisement request, obtains the corresponding keyword, and determines the value type in the advertisement request based on the keyword. Exemplarily, the keyword may be used to perform text matching with the list of advertisements being served, which refers to the list of all advertisements that have been served on the advertisement platform, to determine the value type corresponding to the advertisement request.

[0115] Exemplarily, taking the value types including low value and high value as an example, an exemplary description of the text matching method is given;

[0116] If the keyword has a high degree of text matching with the list of advertisements being served, it can be considered that the value of this advertisement request is high, that is, the value type of this advertisement request is high value; or,

[0117] If the keyword has a low degree of text matching with the list of advertisements being served, it can be considered that the value of this advertisement request is low, that is, the value type of this advertisement request is low value.

[0118] In some alternative embodiments, the manner of determining the value type of the advertisement request may further include: by counting the frequency of the advertisement request.

[0119] For example, if the frequency of the advertisement request for a certain search content is high, the value type of the advertisement request including this search content is high value; or, if the frequency of the advertisement request for a certain search content is low, the value type of the advertisement request including this search content is low value.

[0120] In some alternative embodiments, the manner of determining the value type of the advertisement request may further include: by counting the estimated cost per mille (ECPM) of the advertisement request.

[0121] For example, the value type of the advertisement request with a high ECPM is high value, or the value type of the advertisement request with a low ECPM is low value.

[0122] In some alternative embodiments, the manner of determining the value type of the advertisement request may further include: through the user profile.

[0123] For example, the number of advertisements clicked by the user within a preset time period may be counted. If the number of advertisements clicked by the user is greater than or equal to the threshold, the value type of the advertisement request initiated by this user is high value; or, if the number of advertisements clicked by the user is less than the threshold, the value type of the advertisement request initiated by this user is low value.

[0124] Optionally, step 303 may be executed after step 302:

[0125] Step 303: Use the result of determining the target value type corresponding to the ad request as a virtual ad slot.

[0126] The result of determining the target value type corresponding to the ad request is a virtual ad slot, which includes the target value type corresponding to the determined ad request. The virtual ad slot can be displayed on the ad platform for advertisers to view.

[0127] More specifically, after determining the target value type corresponding to the ad request, compared with the prior art, on the ad platform, "virtual ad slot" information corresponding to the ad request will appear on the display page of the ad request. The virtual ad slot will display the target value type corresponding to the ad request, so that advertisers can intuitively and specifically understand the value of the ad request.

[0128] Table 3 exemplarily shows the virtual ad slot information in the display page.

[0129] Table 3

[0130]

[0131] Referring to Table 3, the display page may include information such as date, real ad slot ID, real ad slot name, virtual ad slot ID, and virtual ad slot name. Among them, the real ad slot is the ad slot visible to the user, that is, the real ad slot is displayed on the second device; the virtual ad slot is the ad slot invisible to the user, that is, the virtual ad slot is not displayed on the second device, but the virtual ad slot can be displayed on the ad platform for advertisers to view.

[0132] Then referring to Table 3, the virtual ad slot information may include the virtual ad slot ID and the virtual ad slot name. Exemplarily, the real ad slot ID "k1f9dhqs06" corresponds to a virtual ad slot with a high value type, and the real ad slot ID "q8oliiiqq5" corresponds to a virtual ad slot with a low value type.

[0133] Optionally, before, during, or after the first device executes step 302, the first device may further execute step 304:

[0134] Step 304: The first device determines the target ad type corresponding to the ad request based on the search content in the ad request.

[0135] In addition to obtaining the target value type of the current ad request, the first device can also obtain the target ad type corresponding to the current ad request.

[0136] In some alternative embodiments, indication information for indicating the advertisement type may be carried in the advertisement request. For example, when the search box corresponds one-to-one to the advertisement type requested by the user, the indication information corresponding to the search box may be carried in the advertisement request. Taking the advertisement types including application advertisements and commodity advertisements as an example, if the user sends an advertisement request through the application advertisement search box, the advertisement request may carry indication information indicating that the requested advertisement this time is an application advertisement; or, if the user sends an advertisement request through the commodity advertisement search box, the advertisement request may carry indication information indicating that the requested advertisement this time is a commodity advertisement.

[0137] In some alternative embodiments, the advertisement type may also be determined by identifying the search content in the advertisement request. For example, by identifying the search content in the advertisement request, the user intention corresponding to the advertisement request can be obtained, and the advertisement attribute information strongly related to it can be obtained through this user intention, where the advertisement attribute information can be used to determine the type of the search advertisement for this search.

[0138] Next, an example is given to illustrate the determination method of keywords and user intention. Suppose the user's search content is: "What are the recommended Double 11 shopping strategies", the keywords that can be obtained through identification are: "Double 11", "shopping"; and the user intention that can be obtained through identification is: shopping. Then, through the user intention, the advertisement attribute information can be determined to be applications related to "shopping".

[0139] Step 305, the first device performs advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request to obtain N search advertisements corresponding to the target value type.

[0140] After obtaining the target value type corresponding to the current advertisement request, advertisement recommendation can be performed based on the target value type of the advertisement request to obtain N search advertisements.

[0141] Wherein, N is a positive integer, and the value of N can be preset according to actual needs. The N search advertisements are the top N search advertisements selected after arranging the N search advertisements corresponding to the target value type corresponding to the current advertisement request in the ranking order of the search advertisements.

[0142] Optionally, if the first device determines the target value type corresponding to the current advertisement request in step 302 and the first device determines the target advertisement type corresponding to the current advertisement request in step 304, that is, when obtaining the target value type and the target advertisement type corresponding to the current advertisement request, step 305 may not be executed, and instead step 306 may be executed.

[0143] Step 306: The first device performs advertisement recommendation for the advertisement request based on the target value type and target advertisement type corresponding to the advertisement request, so as to obtain N search advertisements corresponding to the target value type and target advertisement type.

[0144] In some optional embodiments, the N search advertisements may be the top N search advertisements selected after being arranged in the ranking order of the search advertisements, where the ranking of the N search advertisements may be determined based on ECPM.

[0145] In some optional embodiments, the value of N may be determined according to the size of the search advertisement display area in the second device, where the size information of the search advertisement display area in the second device may be carried in the advertisement request, and the form of the search advertisement display area may specifically refer to Figure 2 the display interface of the second device in, which will not be elaborated here.

[0146] Exemplarily, assume that the search advertisement display area in the second device can accommodate 5 advertisement slots. Then, after the first device performs advertisement recommendation based on the advertisement request, it may send the top 5 advertisements to the second device based on the size of the search advertisement display area indicated by the advertisement request. Or, assume that the search advertisement display area in the second device can accommodate 10 advertisement slots. Then, after the first device performs advertisement recommendation based on the advertisement request, it may send the top 10 advertisements to the second device based on the size of the search advertisement display area indicated by the advertisement request.

[0147] It can be understood that different value types may correspond to different advertisement recommendation models, where different advertisement recommendation models may be used to perform advertisement recommendation for advertisement requests of different values. For example, taking the value types including high value and low value as an example, the advertisement recommendation model may include 2 models, and the 2 models may be a high-value recommendation model and a low-value recommendation model. The high-value recommendation model may be used to perform advertisement recommendation for high-value advertisement requests, and the low-value recommendation model may be used to perform advertisement recommendation for low-value advertisement requests. Thus, advertisement requests of different values can be isolated and decoupled to improve the recommendation effect of search advertisements.

[0148] Step 307: The first device sends the N search advertisements to the second device. Correspondingly, the second device receives the N search advertisements sent by the first device.

[0149] Specifically, when the first device searches for the corresponding N search advertisements based on the advertisement request, it may send these N search advertisements to the second device.

[0150] In some alternative embodiments, when the first device sends the N search ads to the second device, it may also indicate the rankings of the N search ads to the second device. For example, the first device returns the top N search ads.

[0151] Figure 6 Exemplarily, a schematic diagram showing the first device returning the top N search ads is illustrated. Refer to Figure 6 , the top N search ads may be the top N ads selected after sorting according to the ECPM. For example, the top N search ads may include APP_1, APP_2,..., APP_N, where APP_1 has the highest ranking, APP_2 has the second highest ranking, and so on, and APP_N has the lowest ranking.

[0152] Step 308, the second device displays the N search ads.

[0153] Specifically, when the second device receives the N search ads sent by the first device, it may display the N search ads in the search ad display area of the second device.

[0154] In some alternative embodiments, when the second device receives the top N search ads sent by the first device, it may display the top N search ads in the search ad display area of the second device in the ranking order.

[0155] Figure 7 Exemplarily, a schematic diagram showing the second device displaying the top N search ads is illustrated. Refer to Figure 7 , after the second device receives the top N search ads returned by the first device, it may display the top N ads in the search ad display interface in the order from top to bottom. Among them, the higher the search ad is, the higher its ranking. Exemplarily, assuming that APP_1 has the highest ECPM and ranks first, APP_2 has the second highest ECPM and ranks second, and APP_N has the lowest ECPM and ranks last, then APP_1 may be displayed at the top of the search ad display interface, APP_2 may be displayed below APP_1, and APP_N may be displayed at the bottom of the search ad display interface.

[0156] It can be understood that the above-mentioned display method of sequentially displaying in the order from top to bottom according to the ranking of the search ads is only an exemplary illustration and does not constitute a limitation on the embodiments of the present application. It is only a display method set according to the user's visual habit. In some embodiments, it may also be set to other display methods. For example, it may be sequentially displayed in the order from left to right according to the ranking of the search ads.

[0157] In the existing technical solutions for advertisement recommendation, advertisement recommendation is performed on advertisement requests through a unified model, that is, advertisement recommendation is performed through a unified recommendation algorithm. The recommendation result does not distinguish the value of advertisement requests, which easily leads to a low matching degree between the advertisement recommendation result and the user's intention, and a poor advertisement recommendation effect.

[0158] In this application, the corresponding value of an advertisement request is identified, and adaptive advertisement recommendation is performed according to its value to obtain the corresponding search advertisement. Compared with the prior art that does not distinguish the value of advertisement requests, this application can isolate and decouple advertisement requests with different values, thereby improving the recommendation effect of search advertisements, making the recommended search advertisements more in line with the user's intention, and improving the satisfaction of users and advertisers.

[0159] It can be understood that the search advertisements obtained by recommendation through the recommendation models corresponding to different value types can be placed in the corresponding virtual advertisement positions. For example, the search advertisements obtained by performing advertisement recommendation through the recommendation model corresponding to the high-value advertisement type are placed in the corresponding high-value virtual advertisement positions, and the search advertisements obtained by performing advertisement recommendation through the recommendation model corresponding to the low-value advertisement type are placed in the corresponding low-value virtual advertisement positions. Although these virtual advertisement positions are invisible to the end user, they can be shown to advertisers. On this basis, the advertisement platform can set different bidding strategies for different virtual advertisement positions to improve the competitiveness of advertisements and increase the revenue of the advertisement platform.

[0160] As described above through Figure 3 、 Figure 4A 、 Figure 4B 、 Figures 5 - 7 an exemplary description of the advertisement recommendation method is given. Next, the following uses specific embodiments and combines Figure 8 and Figure 9 to give an exemplary description of the advertisement recommendation method.

[0161] Figure 8 FIG. is a schematic diagram of the system architecture of the first device provided in an embodiment of this application. Referring to Figure 8 , the system architecture of the first device may include a search content recognition unit, a value processing unit, and an advertisement recommendation unit.

[0162] Among them, the search content recognition unit can be used to identify the search content in the advertisement request by means of natural language processing (Natural Language Understanding, NLU) and the like to obtain the keywords in the search content.

[0163] In some alternative embodiments, the search content recognition unit can also be used to identify the search content in the advertisement request to determine the advertisement type corresponding to the advertisement request.

[0164] The value processing unit can be used to determine the value type corresponding to the ad request based on keywords. The value type corresponding to the ad request can be divided into high-value type, medium-value type, and low-value type, or can be distinguished according to other value rules.

[0165] It should be noted that the result of determining the value type corresponding to the ad request is a virtual ad slot, that is, the virtual ad slot includes the value type corresponding to the determined ad request. The virtual ad slot can be displayed on the ad platform for advertisers to view.

[0166] The ad recommendation unit can be used to perform corresponding ad recommendations based on the value type to obtain search ads corresponding to the value type.

[0167] In some alternative embodiments, the ad recommendation unit can be implemented through the recommendation algorithm of the Demand-Side Platform (DSP).

[0168] Among them, the DSP can include a recall subunit, a rough ranking subunit, and a fine ranking subunit. The recall subunit can be used to recall candidate ads through a recall algorithm corresponding to the value type. The rough ranking subunit can be used to roughly rank the recalled candidate ads through a rough ranking algorithm corresponding to the value type. The fine ranking subunit can be used to finely rank the ads after rough ranking through a fine ranking algorithm corresponding to the value type to obtain the finely ranked ads. The finely ranked ads can be the top N search ads, and the top N search ads can be used to return to the second device.

[0169] In some alternative embodiments, the first device can also forward the ad request to a third-party DSP and can fuse the result after fine ranking with the output result of the third-party DSP to obtain the top N search ads.

[0170] It can be understood that the DSP corresponding to the ad recommendation unit is the DSP of the ad platform to which the first device belongs, and the third-party DSP can be the DSP of a third-party ad platform.

[0171] Next, on the basis of Figure 8 below, the following combines Figure 9 to give an exemplary description of the generation method of search ads in the first device.

[0172] As Figure 9 shown is a schematic flowchart of another embodiment of the ad recommendation method provided by this application. Steps 302 and 305 can specifically include the following steps:

[0173] Step 901, the first device identifies the search content in the ad request to obtain keywords.

[0174] Specifically, after receiving an advertisement request, the first device can identify the search content in the advertisement request by means of NLU or the like to obtain the keywords corresponding to the search content.

[0175] Among them, the method of obtaining keywords can specifically refer to the relevant descriptions in the above embodiments, which will not be elaborated here.

[0176] Step 902, the first device determines the target value type corresponding to the advertisement request based on the keywords.

[0177] Specifically, when the first device obtains the keywords, it can determine the target value type corresponding to the advertisement request based on the keywords.

[0178] Among them, the way to determine the target value type of the advertisement request can be achieved by text matching of the keywords with the list of advertisements being served. This text matching method can be to match the keywords with the names of all the advertisements being served in the list of advertisements being served.

[0179] Exemplarily, first, the list of advertisements being served can be obtained. Among them, the list of advertisements being served can be obtained by counting all the advertisements being served in the advertisement platform.

[0180] In some alternative embodiments, since the advertisements being served in the advertisement platform are updated dynamically, therefore, the advertisements being served in the advertisement platform can be monitored within a preset time period, and when the advertisements being served in the advertisement platform change, the list of advertisements being served can be updated.

[0181] Then, when the list of advertisements being served is obtained, the keywords can be text-matched with the list of advertisements being served to obtain the value type corresponding to the advertisement request.

[0182] In some alternative embodiments, the way of text matching the keywords with the list of advertisements being served can be a strategy based on brand words. For example, the brand words in the keywords can be obtained, and the brand words can be matched with the text in the list of advertisements being served.

[0183] If the brand word in the keyword matches any advertisement in the list of advertisements being served and the matching degree is relatively high, then the value of this advertisement request is relatively high; or,

[0184] If the brand word in the keyword matches any advertisement in the list of advertisements being served and the matching degree is relatively low, then the value of this advertisement request is relatively low.

[0185] In some alternative embodiments, the text matching results of the keywords with the list of advertisements being served can include multiple results. For example, the text matching results can include: no match, exact match, prefix match, inclusion match, and other multiple results.

[0186] Table 4 exemplarily shows the results of various text matches.

[0187] Table 4

[0188] Search content Results of text matching Matched advertisements "What's the weather like today” Not matched None <![CDATA["Pinduoduo @ ”]]> Precise match <![CDATA[Pinduoduo @ > "Pin” Prefix match <![CDATA[Pinduoduo @ > <![CDATA["How to view the logistics of Pinduoduo" @ > Containment match <![CDATA[Pinduoduo @ >

[0189] Referring to Table 4, for the search content of "How's the weather today", no advertisements of any brand were matched. Therefore, the result of the text match is "not matched". For the search content of "Pinduoduo @ ", an advertisement of the brand "Pinduoduo @ " was matched. Therefore, the result of the text match is "exact match". For the search content of "pin", this search content matches the prefix of the brand "Pinduoduo @ ". Therefore, the result of the text match is "prefix match". For the search content of "How to view Pinduoduo @ logistics", this search content contains the brand "Pinduoduo @ ". Therefore, the result of the text match is "containment match".

[0190] It can be understood that the above results of multiple text matches are only exemplary descriptions and do not constitute a limitation on the embodiments of the present application. In some embodiments, there may also be more or fewer text match results than the above text match results.

[0191] In some alternative embodiments, the result of the text match may form a mapping relationship with the value type. This mapping relationship may be one-to-one, or this mapping relationship may be many-to-one. The embodiments of the present application do not make special limitations on this.

[0192] Exemplarily, taking the results of the text match including multiple results such as not matched, exact match, prefix match, and containment match, and the mapping relationship between the result of the text match and the value type being many-to-one as an example, Table 4 exemplarily shows the mapping relationship between the result of the text match and the value type.

[0193] Table 5

[0194]

[0195] Referring to Table 5, the result of the text match of "not matched" can be mapped to the low-value type, and the results of the text matches of "exact match", "prefix match", and "containment match" can be mapped to the high-value type.

[0196] It can be understood that the above mapping relationship between the result of the text match and the value type is only an exemplary description and does not constitute a limitation on the embodiments of the present application. In some embodiments, the above mapping relationship between the result of the text match and the value type can be set according to actual needs.

[0197] It should be noted that, in some embodiments, the result of the first device determining the value type of the advertisement request based on the keyword is a virtual advertisement space. The virtual advertisement space includes the value types corresponding to the determined advertisement requests. The virtual advertisement space can be displayed on the advertisement platform for advertisers to view, so that advertisers can intuitively perceive the value of advertisement placement.

[0198] Step 903, the first device performs advertisement recommendation on the advertisement request based on the target value type, and obtains N search advertisements corresponding to the target value type.

[0199] Specifically, after the first device obtains the target value type corresponding to the advertisement request, it can perform advertisement recommendation on the advertisement request based on the target value type to obtain N search advertisements corresponding to the target value type.

[0200] It can be understood that for different value types, different recommendation models can be used to recommend advertisement requests of different value types.

[0201] For example, taking the value types including high value and low value as an example, if the value type of the current advertisement request is low value, a recommendation model corresponding to low value can be used for searching to obtain recommended advertisements corresponding to low value; or, if the value type of the current advertisement request is high value, a recommendation model corresponding to high value can be used for recommendation to obtain search advertisements corresponding to high value.

[0202] In some alternative embodiments, the above-mentioned recommendation model can be obtained through pre-training. For example, still taking the value types including high value and low value as an example, the recommendation model corresponding to low value can be trained using sample data containing low value, or the recommendation model corresponding to high value can be trained using sample data containing high value.

[0203] In some alternative embodiments, the recommendation model may include a recall sub-model, a rough ranking sub-model, and a fine ranking sub-model. Among them, at least one of the recall sub-model, the rough ranking sub-model, and the fine ranking sub-model can use a model corresponding to the value type of the current advertisement request.

[0204] For example, the recall sub-model can use a model corresponding to the value type of the current advertisement request. Among them, the recall sub-model can be obtained by pre-training using sample data containing low value and high value, and the rough ranking sub-model and the fine ranking sub-model can use models that have not been trained using sample data containing low value and high value.

[0205] For another example, the recall sub-model and the rough ranking sub-model can use a model corresponding to the value type of the current ad request. Among them, the recall sub-model and the rough ranking sub-model can be obtained by pre-training with sample data including low-value and high-value data. The fine ranking sub-model can use a model that has not been trained with sample data including low-value and high-value data.

[0206] For yet another example, the recall sub-model, the rough ranking sub-model, and the fine ranking sub-model can use a model corresponding to the value type of the current ad request. Among them, the recall sub-model, the rough ranking sub-model, and the fine ranking sub-model can be obtained by pre-training with sample data including low-value and high-value data.

[0207] It can be seen that different ad requests can be isolated and decoupled by different value types, and different ad requests of different value types can be recommended by different recommendation models. Thus, search ads corresponding to different value types can be obtained, thereby improving the recommendation effect and avoiding poor recommendation effects caused by using a unified model to process ad requests with different values.

[0208] Next, taking the recommendation model including a recall sub-model, a rough ranking sub-model, and a fine ranking sub-model as an example, the process of obtaining search ads will be described exemplarily. Assume that the value type of the current ad request is high value, then recommendations can be made through a high-value recall sub-model, a high-value rough ranking sub-model, and a high-value fine ranking sub-model. First, the first device can recall the ads in investment through the high-value recall sub-model to obtain recalled ads. Then, the first device can roughly rank and filter the recalled ads through the high-value rough ranking sub-model to obtain roughly ranked ads. Further, the high-value fine ranking sub-model can finely rank and filter the recalled ads to obtain finely ranked ads.

[0209] It can be understood that each ad in the finely ranked ads has a corresponding predicted click-through rate (pCTR) and predicted conversion rate (pCVR).

[0210] Finally, the finely ranked ads can be sorted by ECPM, and the top N ads in the sorted finely ranked ads can be obtained as search ads.

[0211] Among them, the ECPM of each ad in the finely ranked ads can be calculated through the following formula:

[0212] ECPM = pCTR × pCVR × bid × pid;

[0213] Among them, bid is the bid price of the ad, and pid is the system price adjustment factor.

[0214] For example, assume that 1200 ads are recalled by the high-value recall sub-model. Then, after rough filtering these 1200 recalled ads through the high-value rough ranking sub-model, 200 rough-ranked ads can be obtained. Next, by calculating the ECPM of these 200 rough-ranked ads through the high-value fine ranking sub-model, 200 fine-ranked ads can be obtained. Finally, rank these 200 fine-ranked ads according to the ECPM, and the top N ads can be obtained as the search ads.

[0215] In some alternative embodiments, the top N search ads can also be obtained after ranking the first candidate search ads and the second candidate search ads, where the first candidate search ads can be the candidate ads obtained by the local DSP through the ad recommendation unit for ad recommendation, and the second candidate search ads can be the candidate ads obtained by a third-party DSP for ad recommendation.

[0216] Exemplarily, the first candidate search ads and the second candidate search ads can be fused to form a candidate ad set, where each ad in the candidate ad set has a corresponding ECPM. Then, rank all the ads in the candidate ad set according to the ECPM of each ad, and the top N ads can be obtained. These top N ads are used as the search ads for the user's current request.

[0217] It can be understood that the second candidate search ads can be obtained by the third-party DSP through its own recommendation algorithm for ad recommendation, or the second candidate search ads can be obtained by the third-party DSP through existing or public recommendation algorithms for ad recommendation. The embodiments of the present application do not make special limitations on this.

[0218] Figure 10 This is a schematic structural diagram of an embodiment of the advertisement recommendation device of the present application. As Figure 10 shown, the above advertisement recommendation device 1000 is applied to the first device. The advertisement recommendation device 1000 may include: a receiving module 1010, a determining module 1020, a recommending module 1030, and a sending module 1040; where

[0219] The receiving module 1010 is configured to receive an advertisement request sent by the second device, where the advertisement request includes search content;

[0220] The determining module 1020 is configured to determine the target value type corresponding to the advertisement request based on the search content in the advertisement request;

[0221] The recommending module 1030 is configured to perform advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request, and obtain N search ads corresponding to the target value type, where N is a positive integer;

[0222] A sending module 1040, configured to send the N search ads to the second device.

[0223] In one possible implementation, the above-mentioned determination module 1020 is further configured to use the result of determining the target value type corresponding to the ad request as a virtual ad slot.

[0224] In one possible implementation, the above-mentioned recommendation module 1030 is further configured to determine the target ad type corresponding to the ad request based on the search content in the ad request;

[0225] Based on the target value type and the target ad type corresponding to the ad request, perform ad recommendation for the ad request, and obtain N search ads corresponding to the target value type and the target ad type.

[0226] In one possible implementation, the above-mentioned recommendation module 1030 is further configured to determine the user intention based on the search content in the ad request;

[0227] Determine the target ad type corresponding to the ad request based on the user intention.

[0228] In one possible implementation, the ad types at least include application ads and commodity ads.

[0229] In one possible implementation, the N search ads are the top N search ads, and the ranking of the N search ads is determined by the estimated earnings per thousand impressions ECPM.

[0230] In one possible implementation, the N search ads are obtained after ranking based on the first candidate search ads and the second candidate search ads, where the first candidate search ads are obtained after ad recommendation in the first device, and the second candidate search ads are obtained after ad recommendation in the third device, and the third device is the device of a third-party demand-side platform DSP.

[0231] In one possible implementation, the above-mentioned recommendation module 1030 is further configured to determine a target recommendation model corresponding to the target value type based on the target value type corresponding to the ad request;

[0232] Use the target recommendation model corresponding to the target value type to perform ad recommendation for the ad request, and obtain N search ads corresponding to the target value type.

[0233] In one possible implementation, the above-mentioned determination module 1020 is further configured to identify the search content in the ad request to obtain keywords;

[0234] The keyword is text-matched with the list of advertisements being placed to determine the target value type corresponding to the advertisement request.

[0235] In one possible implementation, the determination module 1020 is further configured to periodically update the list of active advertisements.

[0236] In one possible implementation, the text matching includes multiple results;

[0237] The multiple results are in a one-to-one mapping relationship with the multiple value types; or,

[0238] The multiple results and the multiple value types are in a many-to-one mapping relationship.

[0239] In one possible implementation manner, the advertisement recommendation apparatus 1000 may be a chip or a first device.

[0240] Figure 10 The advertisement recommendation device 1000 provided in the illustrated embodiment may be used to execute the technical solution of the method embodiment illustrated in the present application, and its implementation principle and technical effects may be further referred to the relevant description in the method embodiment.

[0241] It should be understood that the above Figure 10 The division of the various modules of the advertising recommendation device 1000 shown is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software calling through processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software calling through processing elements, and some modules can be implemented in the form of hardware. For example, the detection module can be a separately established processing element, or it can be integrated in a chip of an electronic device. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together, or they can be implemented independently. In the implementation process, each step of the above method or each of the above modules can be completed by an integrated logic circuit of hardware in a processor element or instructions in the form of software.

[0242] For example, the above-mentioned modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), or, one or more Digital Signal Processors (DSPs), or, one or more Field Programmable Gate Arrays (FPGAs), etc. Additionally, these modules may be integrated together and implemented in the form of a System-On-a-Chip (SOC).

[0243] In the above embodiments, the involved processor may include, for example, a CPU, a DSP, a microcontroller or a digital signal processor, and may also include a GPU, an embedded Neural-network Process Unit (NPU), and an Image Signal Processing (ISP). The processor may also include necessary hardware accelerators or logic processing hardware circuits, such as ASICs, or one or more integrated circuits for controlling the execution of the technical solution of this application. In addition, the processor may have the function of operating one or more software programs, and the software programs may be stored in a storage medium.

[0244] The embodiments of this application also provide a readable storage medium, in which a program is stored. When it runs on an electronic device, it causes the electronic device to execute the method provided by the embodiments shown in this application.

[0245] The embodiments of this application also provide a program product, which includes a program. When it runs on an electronic device, it causes the electronic device to execute the method provided by the embodiments shown in this application.

[0246] In the embodiments of this application, "at least one" means one or more, and "a plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent the cases of A existing alone, A and B existing simultaneously, and B existing alone. Here, A and B may be singular or plural. The character " / " generally represents an "or" relationship between the front and rear associated objects. "At least one of the following" and its similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c may represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c may be single or multiple.

[0247] Those of ordinary skill in the art can realize that the units and algorithm steps described in the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0248] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0249] In several embodiments provided by this application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (hereinafter referred to as ROM), random access memories (hereinafter referred to as RAM), magnetic disks, or optical discs that can store program codes.

[0250] The above is only the specific implementation manner of this application. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. The protection scope of this application shall be subject to the protection scope of the claims.

Claims

1. An advertisement recommendation method, characterized in that, Applied to a first device, the method includes: Receiving an advertisement request sent by a second device, where the advertisement request includes search content; Determining a target value type corresponding to the advertisement request based on the search content in the advertisement request; Performing advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request, and obtaining N search advertisements corresponding to the target value type, where N is a positive integer; Sending the N search advertisements to the second device.

2. The method according to claim 1, wherein The method further includes: Taking the result of determining the target value type corresponding to the advertisement request as a virtual advertisement slot.

3. The method according to claim 1 or 2, characterized in that, The performing advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request and obtaining N search advertisements corresponding to the target value type includes: Determining a target advertisement type corresponding to the advertisement request based on the search content in the advertisement request; Performing advertisement recommendation on the advertisement request based on the target value type and the target advertisement type corresponding to the advertisement request, and obtaining the N search advertisements corresponding to the target value type and the target advertisement type.

4. The method according to claim 3, characterized in that, The determining a target advertisement type corresponding to the advertisement request based on the search content in the advertisement request includes: Determining a user intention based on the search content in the advertisement request; Determining the target advertisement type corresponding to the advertisement request based on the user intention.

5. The method according to claim 4, wherein The advertisement types at least include application advertisements and commodity advertisements.

6. The method according to any one of claims 1-5, characterized in that, The N search advertisements are the top N search advertisements, and the ranking of the N search advertisements is determined by the estimated earnings per thousand impressions (ECPM).

7. The method according to claim 6, wherein The N search advertisements are obtained after ranking based on a first candidate search advertisement and a second candidate search advertisement, where the first candidate search advertisement is obtained after advertisement recommendation in the first device, and the second candidate search advertisement is obtained after advertisement recommendation in a third device, and the third device is a device of a third-party demand-side platform (DSP).

8. The method according to any one of claims 1-7, characterized in that, Different value types correspond to different recommendation models; the performing advertisement recommendation on the advertisement request based on the target value type corresponding to the advertisement request and obtaining N search advertisements corresponding to the target value type includes: Determining a target recommendation model corresponding to the target value type based on the target value type corresponding to the advertisement request; Using the target recommendation model corresponding to the target value type to perform advertisement recommendation on the advertisement request, and obtaining N search advertisements corresponding to the target value type.

9. The method according to any one of claims 1-8, characterized in that, The determining a target value type corresponding to the advertisement request based on the search content in the advertisement request includes: Identifying the search content in the advertisement request to obtain keywords; Performing text matching between the keywords and the list of advertisements in progress, and determining the target value type corresponding to the advertisement request.

10. The method according to claim 9, wherein The method further includes: Periodically updating the list of advertisements in progress.

11. The method according to claim 9 or 10, characterized in that, The text matching includes multiple results; The multiple results have a one-to-one mapping relationship with multiple value types; or The multiple results have a many-to-one mapping relationship with multiple value types.

12. An advertising recommendation method, characterized in that, Applied to an advertising recommendation system, the advertising recommendation system includes a first device and a second device, and the method includes: The second device sends an advertising request to the first device in response to search content input by a user; The first device receives the advertising request sent by the second device, where the advertising request includes the search content; The first device determines a target value type corresponding to the advertising request based on the search content in the advertising request; The first device performs advertising recommendation on the advertising request based on the target value type corresponding to the advertising request, and obtains N search advertisements corresponding to the target value type, where N is a positive integer; The first device sends the N search advertisements to the second device; The second device displays the N search advertisements.

13. The method according to claim 12, characterized in that, The method further includes: The first device uses the result of determining the target value type corresponding to the advertising request as a virtual advertising space.

14. The method according to claim 12 or 13, characterized in that, The first device performs advertising recommendation on the advertising request based on the target value type corresponding to the advertising request, and obtains N search advertisements corresponding to the target value type, including: The first device determines a target advertising type corresponding to the advertising request based on the search content in the advertising request; The first device performs advertising recommendation on the advertising request based on the target value type and the target advertising type corresponding to the advertising request, and obtains the N search advertisements corresponding to the target value type and the target advertising type.

15. The method according to claim 14, characterized in that The first device determines a target advertising type corresponding to the advertising request based on the search content in the advertising request, including: The first device determines a user intention based on the search content in the advertising request; The first device determines the target advertising type corresponding to the advertising request based on the user intention.

16. The method according to claim 15, characterized in that, The advertising types at least include application advertisements and commodity advertisements.

17. The method according to any one of claims 12-16, characterized in that, The N search advertisements are the top N search advertisements, and the ranking of the N search advertisements is determined by the estimated earnings per thousand impressions (ECPM).

18. The method according to claim 17, wherein The N search advertisements are obtained after ranking based on a first candidate search advertisement and a second candidate search advertisement, where the first candidate search advertisement is obtained after advertising recommendation in the first device, and the second candidate search advertisement is obtained after advertising recommendation in a third device, and the third device is a device of a third-party demand-side platform (DSP).

19. The method according to any one of claims 12 - 18, characterized in that, Different value types correspond to different recommendation models; the first device performs advertising recommendation on the advertising request based on the target value type corresponding to the advertising request, and obtains N search advertisements corresponding to the target value type, including: The first device determines a target recommendation model corresponding to the target value type based on the target value type corresponding to the advertising request; The first device uses the target recommendation model corresponding to the target value type to perform advertising recommendation on the advertising request, and obtains N search advertisements corresponding to the target value type.

20. The method according to any one of claims 12-19, characterized in that, The first device determines the target value type corresponding to the advertisement request based on the search content in the advertisement request, including: The first device identifies the search content in the advertisement request to obtain keywords; The first device performs text matching between the keywords and the list of advertisements being served to determine the target value type corresponding to the advertisement request.

21. The method according to claim 20, wherein The method further includes: The first device periodically updates the list of advertisements being served.

22. The method according to claim 20 or 21, characterized in that The text matching includes multiple results; There is a one-to-one mapping relationship between the multiple results and multiple value types; or, There is a many-to-one mapping relationship between the multiple results and multiple value types.

23. An advertising recommendation system, characterized in that, Including: A first device and a second device, where the first device is configured to execute any one of the advertisement recommendation methods executed by the first device in claims 12-22, or the first device is configured to execute any one of the advertisement recommendation methods in claims 1-11; the second device is configured to execute any one of the advertisement recommendation methods executed by the second device in claims 12-22.

24. An electronic device, characterized in that, Including: A memory, one or more processors, and one or more programs; wherein, the one or more programs are stored in the memory, and when the one or more processors execute the one or more programs, the electronic device executes the advertisement recommendation method according to any one of claims 1-11.

25. An advertisement recommendation device, characterized in that, Including: One or more functional modules, which are configured to execute the advertisement recommendation method according to any one of claims 1-11.

26. A chip system, characterized in that, The chip system includes a processing circuit and a storage medium, and computer program code is stored in the storage medium; when the computer program code is executed by the processing circuit, the method according to any one of claims 1-22 is implemented.

27. A readable storage medium, characterized in that, The readable storage medium stores a program, and when the program runs on the first device, the advertisement recommendation method according to any one of claims 1-22 is implemented.

Images