Advertising delivery methods, devices, equipment and storage media

Abstract

This application provides an advertising delivery method, apparatus, device, and storage medium, applicable to fields such as advertising. The method includes: obtaining an advertising request from a target object; and based on the advertising request, obtaining content identifiers of N media content consumed by the target object, where the advertising request is used to request an advertisement, and N is a positive integer; based on the content identifiers of the N media content, querying the advertisement identifier corresponding to the content identifier of each of the N media content in a correspondence between content identifiers and advertisement identifiers, where the correspondence includes the advertisement identifier of at least one advertisement corresponding to the content identifier of each of the multiple media content, and the at least one advertisement is an advertisement clicked by at least one object while consuming media content within a preset time period; and delivering the advertisement to the target object based on the advertisement identifier corresponding to the content identifier of each of the N media content, thereby improving the effectiveness of the advertising delivery.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to an advertising delivery method, apparatus, device, and storage medium. Background Technology

[0002] With the continuous development of the internet industry, various media data have become a primary source of information for people. To improve advertising conversion rates, ads can be placed within the media content that the target audience is browsing. For example, ads can be placed within news articles, videos, etc., that the target audience is viewing.

[0003] In the process of advertising, it is necessary to accurately predict the preferences of the target audience (such as users) in order to improve the effectiveness of advertising. However, current advertising methods cannot accurately predict the preferences of the target audience, resulting in unsatisfactory advertising results. Summary of the Invention

[0004] This application provides an advertising delivery method, apparatus, device, and storage medium that can accurately identify the preferences of the target audience, thereby improving the effectiveness of advertising delivery.

[0005] Firstly, this application provides an advertising delivery method, including:

[0006] Obtain the advertising request of the target object, and based on the advertising request, obtain the content identifiers of N media content consumed by the target object, wherein the advertising request is used to request an advertisement, and N is a positive integer;

[0007] Based on the content identifiers of the N media contents, in the correspondence between content identifiers and advertising identifiers, query the advertising identifier corresponding to the content identifier of each of the N media contents. The correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each of the multiple media contents. The at least one advertisement is an advertisement clicked by at least one object while consuming the media contents within a preset time period.

[0008] Based on the advertising identifier corresponding to the content identifier of each of the N media contents, advertisements are delivered to the target audience.

[0009] In some embodiments, determining the normalized value of the ij-th element based on the first numerical value and the second numerical value includes:

[0010] Take the square root of the first value to obtain the third value;

[0011] The ratio of the second value to the third value is determined as the normalized value of the ij-th element.

[0012] In some embodiments, determining the similarity between the k-th media content and each of the T advertisements includes:

[0013] Extract the semantic feature information of the k-th media content;

[0014] Extract the semantic feature information of each of the T advertisements;

[0015] The similarity between the semantic feature information of the k-th media content and the semantic feature information of each of the T advertisements is determined and used as the similarity between the k-th media content and each of the T advertisements.

[0016] Secondly, this application provides an advertising delivery device, comprising:

[0017] The acquisition unit is used to acquire the advertising request of the target object, and based on the advertising request, obtain the content identifiers of N media content consumed by the target object, wherein the advertising request is used to request an advertisement, and N is a positive integer;

[0018] The query unit is used to query the advertising identifier corresponding to the content identifier of each of the N media contents based on the content identifier of the N media contents and in the correspondence between content identifier and advertising identifier. The correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each of the multiple media contents. The at least one advertisement is an advertisement clicked by at least one object while consuming the media contents within a preset time period.

[0019] The processing unit is used to deliver advertisements to the target object based on the advertisement identifier corresponding to the content identifier of each of the N media contents.

[0020] In some embodiments, before obtaining an ad request, the processing unit is further configured to collect P media contents consumed and Q ads clicked by different objects within the preset time period, where P and Q are both positive integers greater than 1; determine the content identifier of each media content in the P media contents, and determine the ad identifier of each ad in the Q ads; associate the content identifiers of the P media contents and the ad identifiers of the Q ads based on the object identifiers to obtain the content identifiers of the media contents consumed and the ad identifiers of the ads clicked by each of M objects within the preset time period, where M is a positive integer greater than 1; and construct the correspondence between the content identifiers and the ad identifiers based on the content identifiers of the media contents consumed and the ad identifiers of the ads clicked by each of the M objects within the preset time period.

[0021] In some embodiments, the processing unit is specifically configured to perform content understanding on the p-th media content among the P media contents to obtain the content identifier of the p-th media content, where p is a positive integer from 1 to P.

[0022] In some embodiments, the processing unit is specifically used to determine the content to be understood corresponding to the p-th media content; to perform content understanding on the content to be understood corresponding to the p-th media content, and to obtain the content identifier of the p-th media content.

[0023] In some embodiments, the processing unit is specifically configured to, if the p-th media content is obtained based on a search request, determine the search keywords in the search request as the content to be understood corresponding to the p-th media content; or, if the p-th media content is not obtained based on a search request, determine the title and summary of the p-th media content as the content to be understood corresponding to the p-th media content.

[0024] In some embodiments, the processing unit is specifically configured to determine the attribute values ​​of K types of attributes of the p-th media content based on the content to be understood, where K is a positive integer; and to perform a hash operation on the attribute values ​​of the K types of attributes to obtain the content identifier of the p-th media content.

[0025] In some embodiments, the processing unit is specifically configured to map the attribute values ​​of each of the K types of attributes to obtain K attribute identifier values; concatenate the K attribute identifier values ​​to obtain concatenated attribute identifier values; and perform a hash operation on the concatenated attribute identifier values ​​to obtain the content identifier of the p-th media content.

[0026] In some embodiments, the processing unit is specifically configured to construct a collaboration matrix based on the content identifier of the media content consumed and the ad identifier of the clicked ads for each of the M objects within the preset time period. The collaboration matrix represents the number of times the M objects click each of the Q ads while consuming each of the P media contents within the preset time period. Based on the collaboration matrix, the unit constructs a correspondence between the content identifier and the ad identifier.

[0027] In some embodiments, the processing unit is specifically configured to determine, based on the content identifiers of the media content consumed and the ad identifiers of the ads clicked by each of the M objects within the preset time period, the number of times each of the P media contents is consumed and each of the Q ads is clicked by each of the M objects within the preset time period; and to construct the collaboration matrix based on the number of times each of the P media contents is consumed and each of the Q ads is clicked by each of the M objects within the preset time period, as well as the content identifiers of the P media contents and the ad identifiers of the Q ads, wherein the ij-th element in the collaboration matrix represents the number of times the M objects consume the ith media content and click the j-th ad, where i is a positive integer less than or equal to P and j is a positive integer less than or equal to Q.

[0028] In some embodiments, the processing unit is specifically used to normalize each element in the collaboration matrix to obtain a normalized collaboration matrix; and to construct the correspondence between the content identifier and the advertisement identifier based on the normalized collaboration matrix.

[0029] In some embodiments, if the collaboration matrix is ​​a P-row, Q-column matrix, the processing unit is specifically configured to, for the ij-th element in the collaboration matrix, obtain the total consumption count of the i-th media content corresponding to the i-th row and the total click count of the j-th advertisement corresponding to the j-th column; determine the normalized value of the ij-th element based on the total consumption count of the i-th media content and the total click count of the j-th advertisement; and obtain the normalized collaboration matrix based on the normalized value of each element in the collaboration matrix.

[0030] In some embodiments, the processing unit is specifically configured to multiply the total number of times the i-th media content is consumed and the total number of times the j-th advertisement is clicked to obtain a first value; obtain from the collaboration matrix the number of times the M objects clicked the j-th advertisement while consuming the i-th media content to obtain a second value; and determine the normalized value of the ij-th element based on the first value and the second value.

[0031] In some embodiments, the processing unit is specifically configured to take the square root of the first value to obtain a third value; and to determine the ratio of the second value to the third value as the normalized value of the ij-th element.

[0032] In some embodiments, the processing unit is specifically configured to, for the k-th media content among P media content, based on the normalized collaboration matrix, select T advertisements from the Q advertisements that have the highest correlation with the k-th media content, where k is a positive integer from 1 to P, the correlation represents the probability value of clicking the advertisement while consuming the k-th media content, and T is a positive integer; determine the similarity between the k-th media content and each of the T advertisements; select S advertisements from the T advertisements that have the highest similarity with the k-th media content, where S is a positive integer less than or equal to T; determine the advertisement identifiers of the S advertisements as the advertisement identifiers of at least one advertisement corresponding to the content identifier of the k-th media content; and construct the correspondence between the content identifier and the advertisement identifier based on the advertisement identifiers of at least one advertisement corresponding to the content identifier of each media content among the P media content.

[0033] In some embodiments, the processing unit is specifically configured to extract semantic feature information of the k-th media content; extract semantic feature information of each of the T advertisements; and determine the similarity between the semantic feature information of the k-th media content and the semantic feature information of each of the T advertisements, as the similarity between the k-th media content and each of the T advertisements.

[0034] Thirdly, this application provides an electronic device including a processor and a memory. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to perform the method described in the first aspect.

[0035] Fourthly, a chip is provided for implementing the methods of various implementations of the first aspect described above. Specifically, the chip includes a processor for retrieving and running a computer program from a memory, causing a device equipped with the chip to perform the methods of the first aspect described above.

[0036] Fifthly, a computer-readable storage medium is provided for storing a computer program that causes a computer to perform the method described in the first aspect.

[0037] In a sixth aspect, a computer program product is provided, including computer program instructions that cause a computer to perform the method described in the first aspect.

[0038] In a seventh aspect, a computer program is provided that, when run on a computer, causes the computer to perform the method described in the first aspect.

[0039] In summary, this application constructs a correspondence between content identifiers and advertising identifiers. This correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each media content in multiple media content sets. This at least one advertisement is an advertisement clicked by at least one object while consuming the media content within a preset time period. Thus, when obtaining an advertising request from a target object, the content identifiers of N media content sets consumed by the target object are obtained based on the advertising request. Then, based on these N media content content identifiers, the advertising identifier corresponding to the content identifier of each of the N media content sets is queried from the correspondence between content identifiers and advertising identifiers. Finally, based on the advertising identifier corresponding to the content identifier of each of the N media content sets, the advertisement is delivered to the target object. Therefore, this application embodiment first constructs a correspondence between content identifiers and advertising identifiers based on the consumption behavior of different objects towards different media content and the clicking behavior of different objects towards different advertisements. This correspondence, including the advertising identifier of at least one advertisement corresponding to the content identifier of each media content in multiple media content sets, can accurately reflect the advertisements clicked by objects while consuming media content. This allows for the retrieval of advertising identifiers corresponding to the content identifiers of N media content consumed by the target audience within a specific mapping. This enables accurate prediction of advertisements that the target audience is interested in, reduces the target audience's aversion to advertisements, and improves the effectiveness of ad placement. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1A and Figure 1B A schematic diagram illustrating existing advertising recall methods.

[0042] Figure 2 A schematic diagram illustrating the implementation environment of an advertising delivery method provided in this application embodiment;

[0043] Figure 3 A flowchart illustrating an advertising delivery method provided in an embodiment of this application;

[0044] Figure 4 This is a schematic diagram illustrating the key steps of the advertising recall involved in the embodiments of this application;

[0045] Figure 5AA diagram illustrating the content identifiers of consumed media content and the ad identifiers of clicked ads for each object within a preset time period;

[0046] Figure 5B This is a schematic diagram of a coordination matrix;

[0047] Figure 5C A schematic diagram of the normalized synergy matrix;

[0048] Figure 6 A flowchart illustrating an advertising delivery method provided in an embodiment of this application;

[0049] Figure 7 A schematic diagram of the framework of an advertising delivery method provided in an embodiment of this application;

[0050] Figure 8 This is a schematic block diagram of an advertising delivery device provided in one embodiment of this application;

[0051] Figure 9 This is a schematic block diagram of the electronic device provided in the embodiments of this application. Detailed Implementation

[0052] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0053] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. In embodiments of the invention, "B corresponding to A" means that B is associated with A. In one implementation, B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0054] The advertising delivery method provided in this application can be applied to various fields such as advertising delivery and media data processing. During the advertising delivery process, it can accurately predict the advertisements that the target audience is interested in based on the target media content requested by the target audience, thereby improving the advertising delivery effect.

[0055] To facilitate understanding of the embodiments of this application, the relevant concepts involved in the embodiments of this application will first be introduced:

[0056] Advertising placement involves using various formats such as text, images, or videos to deliver ads to a precise target audience based on their needs, attracting their attention and clicks, thereby converting them into potential customers and ultimately closing the deal.

[0057] Recall Model: In a recommendation system, based on the user's feature requests, a certain number of ads are recalled from the full ad library to serve as an initial filter for ad volume, which is then passed on to coarse and fine ranking.

[0058] Collaborative filtering is the most basic recommendation method, making recommendations based on similar users or similar items. It achieves the goal of recommendation by grouping people according to similarities and items according to similarity.

[0059] Iceberg: Iceberg tables are an open-source data table format used for storing and managing large-scale datasets in cloud storage. It aims to provide functionality similar to traditional database management systems, such as transaction support, schema evolution, and data version control. Iceberg tables are commonly used in big data processing frameworks such as Apache Spark and Apache Hive to provide more reliable and efficient data management and querying capabilities.

[0060] Current advertising delivery methods mainly include user-based collaborative filtering (i.e., grouping people by similar traits) and ad-based collaborative filtering (i.e., clustering similar items). Among these, user-based collaborative filtering methods include... Figure 1A As shown, a user similarity matrix is ​​constructed (where A, B, C, and D represent four different users, and ad represents an advertisement). Based on the principle that similar users have similar interests and therefore similar advertisements, advertisements with similar preferences are recommended to users. For example, if user A and user B are similar users, and user A clicks on advertisement 1, then user B will also be shown advertisement 1. This is an example of an ad-based collaborative filtering method for ad delivery. Figure 1BAs shown, a similarity matrix is ​​constructed between ads. It's assumed that if two ads have been clicked by many people, they are considered similar, and similar ads that have been clicked are recommended to users. For example, if ad 1 and ad 2 are similar, and user A clicks on ad 1, ad 2 is recommended to user A. However, because ads are highly targeted, it's difficult to provide a good user experience. A user clicking on an ad doesn't necessarily mean they like the recommended ad. If the recommended ad doesn't cause offense, it's already largely successful. Therefore, in ad recommendation scenarios, recommending ads that another user has clicked based on similar user preferences isn't necessarily a preferred method. Thus, current ad delivery methods, due to their inability to accurately predict target audience preferences, lead to unpredictable ad delivery effectiveness.

[0061] To address the aforementioned technical problems, this application proposes a novel advertising delivery method. First, a correspondence between content identifiers and advertising identifiers is constructed. This correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each media content across multiple media content sets. This at least one advertisement is an advertisement clicked by at least one object while consuming the media content within a preset time period. Thus, when obtaining an advertising request from a target object, the content identifiers of N media content sets consumed by the target object are obtained based on the advertising request. Then, based on the content identifiers of these N media content sets, the advertising identifier corresponding to the content identifier of each of the N media content sets is queried from the correspondence between content identifiers and advertising identifiers. Finally, based on the advertising identifiers corresponding to the content identifiers of these N media content sets, advertising is delivered to the target object. Therefore, this application embodiment first constructs a correspondence between content identifiers and advertising identifiers based on the consumption behavior of different objects towards different media content and the clicking behavior of different objects towards different advertisements. This correspondence, including the advertising identifier of at least one advertisement corresponding to the content identifier of each media content across multiple media content sets, can accurately reflect the advertisements clicked by objects while consuming media content. This allows for the retrieval of advertising identifiers corresponding to the content identifiers of N media content consumed by the target audience within a specific mapping. This enables accurate prediction of advertisements that the target audience is interested in, reduces the target audience's aversion to advertisements, and improves the effectiveness of ad placement.

[0062] The implementation environment of the embodiments of this application is described below.

[0063] Figure 2 A schematic diagram illustrating the implementation environment of an advertising delivery method provided in this application embodiment, as shown below. Figure 2As shown, this implementation environment includes: terminal device 110, server 120, and advertising platform 130. Terminal device 110 and server 120 are connected via wired or wireless means. Advertising platform 130 is connected to server 120 via wired or wireless means.

[0064] In this embodiment, the user can consume media content through the terminal device 110, such as watching videos, reading articles, and searching for content of interest. Simultaneously, the user can also view advertisements through the terminal device 110.

[0065] In this embodiment, server 120 is mainly used to provide data to terminal device 110, process various requests sent by terminal device 110, and send the processing results of the requests to terminal device 110. For example, server 120 provides media content data to terminal device 110, provides advertising content to terminal device 110, etc.

[0066] The advertising delivery method provided in this application first constructs a correspondence between content identifiers and advertising identifiers. This correspondence includes an advertising identifier for at least one advertisement corresponding to the content identifier of each media content in multiple media contents. The at least one advertisement is an advertisement clicked by at least one object while consuming media content within a preset time period.

[0067] The advertising delivery method provided in this application embodiment can be performed by the terminal device 110, the server 120, or a system composed of the terminal device 110 and the server 120. This application embodiment does not impose any limitations on this.

[0068] In some embodiments, the advertising delivery method of this application is performed by server 120. In this case, as... Figure 2As shown, terminal device 110 sends an advertising request to server 120 for the target object. For example, when the target object views or is about to view an advertising display position, terminal device 110 sends an advertising request to server 120. This advertising request is used to request an advertisement. After receiving the advertising request sent by terminal device 110, server 120 obtains the content identifiers of N media content consumed by the target object in the recent time period based on the advertising request. Next, based on the content identifiers of the N media content, server 120 queries the correspondence between content identifiers and advertising identifiers to find the corresponding advertising identifier for each content identifier among the N media content. The correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each media content. The at least one advertisement is an advertisement clicked by at least one object while consuming media content within a preset time period. Finally, server 120 delivers the advertisement to the target object based on the advertising identifier corresponding to the content identifier of each media content among the N media content. For example, server 120 requests an advertisement from advertising platform 130 based on the advertising identifier corresponding to the content identifier of each media content among the N media content. Advertising platform 130 sends advertising data for at least one advertisement corresponding to the advertising identifier of each of the N media contents to server 120. Server 120 then sends the advertising data for at least one advertisement to terminal device 110. Terminal device 110 displays the at least one advertisement to the target audience.

[0069] Therefore, this embodiment first constructs a correspondence between content identifiers and ad identifiers based on the consumption behavior of different objects towards different media content and their click behavior towards different advertisements. This correspondence includes the ad identifier of at least one advertisement corresponding to the content identifier of each media content, accurately reflecting the advertisements clicked by the object when consuming media content. Thus, based on the content identifiers of N media content consumed by the target object, the corresponding ad identifiers can be retrieved from this correspondence, enabling accurate prediction of advertisements that the target object is interested in, reducing the target object's aversion to advertisements, and improving the effectiveness of ad placement.

[0070] In some embodiments, the terminal device 110 includes, but is not limited to, desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices may include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle systems. Portable wearable devices may include smartwatches, smart bracelets, and head-mounted devices. Terminal devices are often equipped with a display device, which may be a monitor, display screen, touchscreen, etc., and the touchscreen may be a touchscreen, touch panel, etc.

[0071] In some embodiments, the server 120 or advertising platform 130 may be one or more servers. When there are multiple servers, at least two servers may be used to provide different services, and / or at least two servers may be used to provide the same service, such as providing the same service in a load-balanced manner. This application embodiment does not limit this. The server may be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. A server may also be a node in a blockchain.

[0072] It should be noted that the implementation environment of this application embodiment includes, but is not limited to, Figure 2 As shown.

[0073] The technical solutions of the embodiments of this application will be described in detail below through some examples. The following embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0074] Figure 3 This is a schematic flowchart illustrating an advertising delivery method according to an embodiment of this application. The execution entity of this embodiment can be a device with advertising delivery functionality, such as an advertising delivery device. For example, the advertising delivery device can be the one described above. Figure 2 The terminal device shown may be a server or a system consisting of a terminal device and a server. For ease of description, the following embodiments use an electronic device as an example to illustrate the advertising delivery method of this application.

[0075] like Figure 3 As shown, the advertising delivery method in this application embodiment includes the following steps:

[0076] S101. Obtain the advertising request of the target object, and based on the advertising request, obtain the content identifiers of the N media content consumed by the target object.

[0077] Here, the ad request is used to request an ad, and N is a positive integer.

[0078] The target object in this application embodiment can be understood as any object consuming media content. For example, an object reading an article, an object watching a video, or an object performing a data search, etc.

[0079] In this embodiment, advertisements can be served to the target audience while they consume media content. For example, advertisements can be served to the target audience while they are reading an article, watching a video, or performing a data search.

[0080] In this embodiment of the application, when delivering an advertisement to a target object, the advertising request of the target object is first obtained. This advertising request is used to request an advertisement. For example, if the target object is browsing media data (e.g., reading resources, or watching movies / TV shows) on a media platform through a terminal device, an advertisement needs to be displayed in the advertising area. At this time, the terminal device will initiate an advertising request.

[0081] This application does not limit the specific method by which an electronic device obtains advertising requests from a target object.

[0082] In some embodiments, if the electronic device is a terminal device, when the target user encounters an ad slot while consuming media content, the terminal device generates an ad request for the target user. For example, when the target user is watching a video on a video platform and needs to load ads after watching several videos, the terminal device generates an ad request for the target user. In this way, the terminal device can obtain the generated ad request.

[0083] In some embodiments, if the electronic device is a server, when the target user encounters an advertisement while consuming media content, the terminal device generates an advertisement request for the target user. The terminal device then sends the generated advertisement request to the server. This allows the server to obtain the advertisement request from the target user.

[0084] In this embodiment of the application, after the electronic device obtains the advertising request of the target object, it obtains the content identifiers of N media content consumed by the target object based on the advertising request.

[0085] It should be noted that, in the process of constructing the object relationship between content identifiers and advertising identifiers in this application embodiment, the content identifiers of the media content consumed by different objects in a recent period are determined. This process is described in detail in the following embodiments and will not be repeated here. For example, the content identifiers of the media content consumed by different objects in a recent period are shown in Table 1:

[0086] Table 1

[0087]

[0088] In Table 1 above, Uid represents the object identifier, and content_id represents the content identifier of the media content. As shown in Table 1, the content identifiers of the media content consumed by object Uid1 in the recent period are content_id_1, content_id_2, etc. It should be noted that Table 1 is updated in real time, for example, every certain period of time, and only the content identifiers of the media content recently consumed by different objects are retained. In this way, during this advertising campaign, the content identifiers of N media content consumed by the target object in the recent period can be obtained.

[0089] This application embodiment does not limit the specific method by which the electronic device obtains the content identifiers of the N media contents consumed by the target object based on the advertising request.

[0090] In some possible implementations, the advertising request includes the object identifier of the target object. Thus, after receiving the advertising request, the electronic device parses it to obtain the object identifier of the target object, and then, based on the object identifier, looks up the content identifiers of the N media content items consumed by the target object in Table 1.

[0091] In one possible implementation, the ad request includes content identifiers of N media content items already consumed by the target object. For example, when a terminal device sends an ad request to the server, it queries Table 1 based on the target object's object identifier to obtain the content identifiers of the N media content items already consumed by the target object, and then includes these content identifiers in the ad request.

[0092] In one possible scenario, the N media content items consumed by the target object can be identified by the content identifiers of all media content consumed by the target object in the recent period, as stored in Table 1 above. That is, the number of media content items consumed in the recent period may differ for different objects, and thus the corresponding N may also differ.

[0093] In one possible scenario, the N media contents consumed by the target object can be the content identifiers of the N media contents consumed by the target object in the recent period, as stored in Table 1 above. That is, the number of N media contents selected in this embodiment is the same for different objects.

[0094] In this embodiment of the application, after the electronic device obtains the content identifiers of the N media contents consumed by the target object based on the above steps, it executes the following step S102.

[0095] S102. Based on the content identifiers of N media contents, in the correspondence between content identifiers and advertising identifiers, query the advertising identifier corresponding to the content identifier of each of the N media contents.

[0096] The correspondence includes the ad identifier of at least one advertisement corresponding to the content identifier of each media content in multiple media content, and the at least one advertisement is the advertisement clicked by at least one object while consuming media content within a preset time period.

[0097] In this embodiment, before obtaining the advertising request of the target object, it is first necessary to construct a correspondence between content identifiers and advertising identifiers. This correspondence can be constructed by the aforementioned electronic device or by other devices. The electronic device simply loads the constructed correspondence. For ease of description, the following uses the example of an electronic device constructing the correspondence between content identifiers and advertising identifiers to illustrate the specific process.

[0098] In some embodiments, the electronic device can construct the correspondence between content identifiers and advertisement identifiers through the following steps A and B:

[0099] Step A: Collect the P media content items consumed and the Q advertisements clicked by different objects within a preset time period, where P and Q are both positive integers greater than 1;

[0100] Step B: Determine the content identifier for each of the P media contents and the ad identifier for each of the Q ads;

[0101] Step C: Based on the object identifier, associate the content identifiers of P media contents with the ad identifiers of Q advertisements to obtain the content identifiers of consumed media contents and the ad identifiers of clicked advertisements for each of the M objects within a preset time period, where M is a positive integer greater than 1.

[0102] Step D: Based on the content identifiers of consumed media content and the ad identifiers of clicked ads for each of the M objects within a preset time period, construct the correspondence between content identifiers and ad identifiers.

[0103] The aforementioned preset time period can be understood as the period closest to the current moment, such as the 10 days or 7 days before the current moment.

[0104] In this embodiment of the application, when the electronic device completes the correspondence between content identifiers and advertising identifiers, it obtains multiple objects, such as the behavior of M objects in the content domain and the behavior in the advertising domain.

[0105] In this context, an object's behavior in the content domain can be understood as the object's behavior of consuming media content, such as the object reading articles, watching videos, and searching data.

[0106] In this context, an object's behavior in the advertising domain can be understood as the object clicking on an advertisement. For example, the scenarios corresponding to the advertising domain include advertising scenarios within the content feed, such as an object clicking on an advertisement after reading to the bottom of a WeChat official account, or an object clicking on a video advertisement. Optionally, the scenarios corresponding to this advertising domain also include other scenarios, such as advertising scenarios in WeChat Moments.

[0107] For example, such as Figure 4 As shown in the embodiments of this application, the electronic device collects the media content consumed by each object within a preset time period based on the behavior of each object in the content domain.

[0108] For example, Table 2 shows the media content consumed by each of the different objects collected by the electronic device within a preset time period:

[0109] Table 2

[0110] Object A Media Content 1, Media Content 2, Media Content 3 Object B Media Content 1 Object C Media Content 1, Media Content 2 Object D Media Content 2 …… ……

[0111] As shown in Table 2, the media content consumed by different objects within a preset time period is denoted as P media contents. For example, if the media content consumed by objects A, B, C, and D within the preset time period is media content 1, media content 2, and media content 3, then P equals 3. In other words, the aforementioned P media contents represent the different media contents consumed by the objects within the preset time period.

[0112] At the same time, based on the behavior of each object in the advertising domain, the electronic device collects the ads that each object has clicked within a preset time period.

[0113] For example, Table 3 shows the ads that have been clicked by each of the different objects collected by the electronic device within a preset time period:

[0114] Table 3

[0115] Object A Advertisement 1, Advertisement 2 Object B Advertisement 1, Advertisement 3 Object C Advertisement 2, Advertisement 3 Object D Advertisement 1, Advertisement 3 …… ……

[0116] As shown in Table 3, the electronic device collects Q advertisements that different objects have clicked within a preset time period. For example, for objects A, B, C, and D, the advertisements they have clicked within the preset time period are advertisement 1, advertisement 2, and advertisement 3, which are three different advertisements. In this case, Q equals 3. That is to say, the above Q advertisements are the different advertisements that the objects have clicked within the preset time period.

[0117] Next, the electronic device determines the content identifier of each of the P media contents and the ad identifier of each of the Q ads.

[0118] In this embodiment, the advertisement ID can be directly identified as the advertisement identifier. The advertisement ID is an inherent attribute of the advertisement and can be directly obtained from the advertisement data. For example, Figure 4 As shown, the ad identifier for an ad clicked by a certain object is: uid->adgroup_id, where uid is the object identifier and adgroup_id is the ad identifier.

[0119] The following describes the specific process of determining the content identifier for each of the P media contents.

[0120] In the embodiments of this application, the specific methods for determining the content identifier of each media content in P media content are basically the same. For ease of description, the specific process of determining the content identifier of one media content in P media content, such as the p-th media content, will be described here.

[0121] This application does not limit the specific method by which the electronic device determines the content identifier of the p-th media content.

[0122] In some embodiments, the ID of the p-th media content can be determined as the content identifier of the p-th media content, where the ID of the p-th media content is used to uniquely identify the p-th media content. For example, each article on a WeChat official account has a unique identifier ID. If the p-th media content is an article on a WeChat official account, then the unique identifier ID of the p-th media content on the WeChat official account can be determined as the content identifier of the p-th media content.

[0123] In some embodiments, the electronic device may determine the content identifier of the p-th media content through the following steps B1 and B2:

[0124] Step B1: Determine the content to be understood corresponding to the p-th media content;

[0125] Step B2: Perform content understanding on the content to be understood corresponding to the p-th media content to obtain the content identifier of the p-th media content.

[0126] In this implementation, the electronic device determines the content to be understood corresponding to the p-th media content, and then performs content understanding on the content to be understood corresponding to the p-th media content to obtain the content identifier of the p-th media content. For example, the electronic device calls a content understanding service to map different media content to a unified dimensional space to obtain the content identifier of each media content. For example, the content identifier of the media content consumed by a certain object can be recorded as: uid->content_id, where uid is the object identifier and content_id is the advertisement identifier.

[0127] This application does not limit the specific method by which the electronic device determines the content to be understood corresponding to the p-th media content. In this application, the content to be understood can be customized for different scenarios.

[0128] In one example, if the p-th media content is obtained based on a search request, then the search keywords in the search request are identified as the content to be understood corresponding to the p-th media content. For example, in a reading article scenario, where the p-th media content is an article, then the title and author description of the p-th media content are identified as the content to be understood for the p-th media content. In a search scenario, where the p-th media content is obtained based on a search request input by an object, then the search keywords input by the object are identified as the content to be understood for the p-th media content.

[0129] In one example, if the p-th media content was not obtained through a search request, then the title and summary of the p-th media content are identified as the content to be understood. For instance, in an article reading scenario, where the p-th media content is an article, the title and author's description (i.e., the summary) are identified as the content to be understood. Similarly, in a video watching scenario, where the p-th media content is a video, the title and summary are identified as the content to be understood, where the summary can be descriptive information entered by the author.

[0130] Based on the above steps, after determining the content to be understood for the p-th media content, the electronic device performs content understanding on the content to be understood corresponding to the p-th media content to obtain the content identifier of the p-th media content.

[0131] This application embodiment does not limit the specific method by which the electronic device performs content understanding on the content to be understood corresponding to the p-th media content to obtain the content identifier of the p-th media content.

[0132] In some embodiments, the electronic device loads a content identifier generation model, which can generate a content identifier for the input media content based on the content to be understood. Thus, the electronic device can input the content to be understood corresponding to the p-th media content into the content identifier generation model and output the content identifier for the p-th media content.

[0133] In some embodiments, based on the content to be understood corresponding to the p-th media content, the attribute values ​​of K types of attributes of the p-th media content are determined, where K is a positive integer. These K types of attributes are K preset attributes of different types, such as category, brand, and business concept attributes. Thus, the electronic device performs content understanding on the content to be understood corresponding to the p-th media content to obtain the attribute values ​​of the category, brand, and business concept attributes of the p-th media content, i.e., to obtain the specific category, specific brand, and specific business concept of the p-th media content. Next, a hash operation is performed on the attribute values ​​of the K types of attributes of the p-th media content to obtain the content identifier of the p-th media content.

[0134] The methods for hashing the attribute values ​​of the K-type attributes of the p-th media content to obtain the content identifier of the p-th media content include at least the following:

[0135] Method 1: Extract the feature vector of each attribute value from the K attribute values ​​of the p-th media content. Next, concatenate the feature vectors of the K attribute values ​​to obtain a concatenated vector. Perform a hash operation on this concatenated vector to obtain a hash value, which is then used as the content identifier for the p-th media content.

[0136] Method 2 involves mapping the attribute values ​​of each of the K categories of attributes to their corresponding IDs to obtain K attribute identifiers; concatenating the K attribute identifiers to obtain a concatenated attribute identifier; and performing a hash operation on the concatenated attribute identifier to obtain the content identifier of the p-th media content.

[0137] In one example of Method 2, an ID is assigned to each possible attribute value of each of the K categories of attributes. Thus, for the attribute value 'a' of attribute i in the K categories of the p-th media content, the ID corresponding to attribute value 'a' is searched among the IDs of all possible attribute values ​​under attribute i, and this ID is determined as the attribute identifier value corresponding to attribute value 'a'. This yields the IDs corresponding to the attribute values ​​of each of the K categories of attributes in the p-th media content, resulting in K attribute identifier values. Next, the K attribute identifier values ​​(i.e., the K IDs) are concatenated and hashed to output a single ID value, which is recorded as the content identifier 'content_id' of the p-th media content.

[0138] Following the steps described above, the electronic device can determine the content identifier of each of the P media contents consumed by different objects within a preset time period.

[0139] In this embodiment of the application, the above steps can be used to obtain the behavior of the object in the advertising domain and the behavior of the object in the content domain, thereby generating data in the format shown in the example below:

[0140] For example, the content identifier of the media content consumed by an object, determined based on the object's search behavior, can be represented as: uid->content_id.

[0141] For example, the content identifier of the media content consumed by an object, determined based on the object's article reading behavior, can be represented as: uid->appid->content_id, where appid can be understood as the software identifier of the object reading the article.

[0142] For example, the ad identifier of an ad that an object clicks on can be represented as: uid->adgroup_id.

[0143] Next, the electronic device performs step C above, such as... Figure 4 As shown, based on the object identifier, the content identifiers of P media contents and the ad identifiers of Q advertisements are associated to obtain the content identifiers of consumed media contents and the ad identifiers of clicked advertisements for each of the M objects within the preset time period.

[0144] For example, suppose the P media contents mentioned above are Media Content 1, Media Content 2, Media Content 3, and Media Content 4, and suppose the Q advertisements mentioned above are Advertisement 1, Advertisement 2, and Advertisement 3. Assuming that the advertisements clicked by objects A, B, C, and D within a preset time period are as shown in Table 2 above, the correspondence between the object identifiers of these four objects and the advertisement identifiers of the three advertisements can be determined as shown in Table 4.

[0145] Table 4

[0146] uid_A Object A Advertisement 1, Advertisement 2 adgroup_id_1, adgroup_id_2 uid_B Object B Advertisement 1, Advertisement 3 adgroup_id_1, adgroup_id_3 uid_C Object C Advertisement 2, Advertisement 3 adgroup_id_2, adgroup_id_3 uid_D Object D Advertisement 1, Advertisement 3 adgroup_id_1, adgroup_id_3 …… …… …… ……

[0147] Assuming that the media content consumed by four objects—A, B, C, and D—within a preset time period is as shown in Table 3 above, the correspondence between the object identifiers of these four objects and the content identifiers of the three media contents can be determined as shown in Table 5.

[0148] Table 5

[0149]

[0150] In this way, electronic devices can determine the content identifiers of consumed media content and the ad identifiers of clicked advertisements for each of M objects within a preset time period, based on the object identifiers. For example, for object A, by combining Tables 4 and 5 above based on uid_A, it can be determined that the content identifiers of consumed media content for object A within the preset time period are content_id_1, content_id_2, and content_id_3, and the ad identifiers of clicked advertisements are adgroup_id_1 and adgroup_id_2.

[0151] In one example, for each of the M objects, the content identifiers of the consumed media content and the ad identifiers of the clicked ads within a preset time period are shown in Table 6. Figure 5A As shown:

[0152] Table 6

[0153]

[0154] In the embodiments of this application, as shown in Table 6 above and Figure 5A As shown, by associating an object's behavior in the content domain and its behavior in the ad domain using the object ID, a combination of an object's actions of simultaneously clicking an ad and browsing / searching content can be obtained. Figure 5A In this code, A, B, C, and D are abbreviations for object identifiers uid_A, uid_B, uid_C, and uid_D; C1, C2, and C3 are abbreviations for content identifiers content_id_1, content_id_2, and content_id_3; and ad1, ad2, and ad3 are abbreviations for advertising identifiers adgroup_id_1, adgroup_id_2, and adgroup_id_3.

[0155] Next, the electronic device performs step D above, constructing a correspondence between content identifiers and ad identifiers based on the content identifiers of consumed media content and the ad identifiers of clicked advertisements for each of the M objects within the preset time period.

[0156] This application embodiment does not limit the specific method by which the electronic device constructs the correspondence between content identifiers and ad identifiers based on the content identifiers of consumed media content and the ad identifiers of clicked advertisements for each of the M objects within the preset time period.

[0157] In some embodiments, the electronic device, based on the content identifier of the media content consumed and the ad identifier of the clicked ads for each of the M objects within the preset time period, counts each of the P media contents, and when the M objects consume the media content, clicks the few ads that are clicked the most times among the Q ads, and then matches the content identifier of the media content with the ad identifier of the few ads.

[0158] In some embodiments, the electronic device constructs a mapping between content identifiers and advertisement identifiers through the following steps:

[0159] Step D1: Based on the content identifier of the media content consumed and the ad identifier of the clicked ads for each of the M objects within a preset time period, construct a collaboration matrix. This collaboration matrix represents the number of times each of the Q ads is clicked while each of the P media contents is consumed by the M objects within the preset time period.

[0160] Step D2: Based on the collaboration matrix, construct the correspondence between content identifiers and advertising identifiers.

[0161] In this implementation, the electronic device constructs a collaboration matrix based on the content identifiers of the media content consumed and the ad identifiers of the clicked ads for each of the M objects within a preset time period. This collaboration matrix represents the number of times each of the Q ads is clicked while each of the P media contents is consumed by the M objects within the preset time period.

[0162] The embodiments of this application do not limit the specific form of the cooperative matrix.

[0163] In one example, the collaboration matrix is ​​a Q-row, P-column matrix, where each of the Q rows represents the ad identifier of one of the Q ads, and each of the P columns represents the content identifier of one of the P media content pieces. The qp-th element of the collaboration matrix represents the number of times M objects consume the p-th media content while clicking the q-th ad.

[0164] In one example, based on the content identifiers of the media content consumed and the ad identifiers of the ads clicked by each of the M objects within a preset time period, the number of times each of the P media contents consumed and each of the Q ads clicked by the M objects within the preset time period is determined. Based on the number of times each of the P media contents consumed and each of the Q ads clicked by the M objects within the preset time period, as well as the content identifiers of the P media contents and the ad identifiers of the Q ads, a collaboration matrix is ​​constructed. The ij-th element in the collaboration matrix represents the number of times the M objects consumed the ith media content and clicked the j-th ad, where i is a positive integer less than or equal to P and j is a positive integer less than or equal to Q.

[0165] It should be noted that, in this embodiment, clicking on an advertisement while consuming media content is not limited to the act of clicking on the advertisement during the process of consuming the media content. It also includes the act of clicking on the advertisement after consuming the media content, or the act of clicking on the advertisement before consuming the media content. As long as the time when the object consumes the media content and the time when the object clicks on the advertisement are within a preset time period, it can be understood that the object clicked on the advertisement while consuming the media content.

[0166] For example, assuming that for each of the M objects within a preset time period, the content identifiers of the consumed media content and the ad identifiers of the clicked ads are as shown in Table 6 above. Figure 5A As shown in Table 6, based on the content identifiers of the media content consumed and the ad identifiers of the clicked ads for each of the P media contents consumed within a preset time period, the number of times each of the Q ads clicked by each of the M objects within the preset time period is determined. For example, objects A, B, C, and D, objects A and C both clicked ad 1 while consuming media content 1. If the M objects are these four objects, then it can be determined that these M objects clicked ad 1 2 times while consuming media content 1. Based on this, according to Table 6 above, the number of times each of the Q ads clicked by each of the M objects within the preset time period while consuming each of the P media contents is determined, as shown in Table 7 below. Figure 5B As shown:

[0167] Table 7

[0168] adgroup_id_1 adgroup_id_2 adgroup_id_3 content_id_1 2 2 2 content_id_2 1 2 1 content_id_3 1 0 0 …… …… ……

[0169] As shown in Table 7, the electronic device removes the object identifiers from Table 6 above, aggregates the content identifiers and advertising identifiers, and thus obtains the number of times M objects consume each of the media content in P media content and click on each of the advertisements in Q advertisements within the preset time period.

[0170] In this way, the electronic device constructs a collaboration matrix based on the number of times each of the P media contents consumed by M objects within a preset time period, and the content identifiers of the P media contents and the ad identifiers of the Q ads. For example, Table 7 above can be defined as the collaboration matrix, where the element values ​​are the co-occurrence counts of ad identifiers and content identifiers. Here, the co-occurrence count can be represented by a_c_cnt, indicating that the object both clicked on an ad and searched or browsed media content. For example, the ij-th element in this collaboration matrix represents the number of times the M objects consumed the i-th media content and clicked the j-th ad.

[0171] After determining the collaboration matrix based on the above steps, the electronic device constructs the correspondence between content identifiers and advertising identifiers based on the collaboration matrix.

[0172] This application does not limit the specific method by which electronic devices construct the correspondence between content identifiers and advertising identifiers based on the collaborative matrix.

[0173] In one possible implementation, the aforementioned collaboration matrix represents the number of times each of the Q advertisements is clicked when M objects consume each of the P media contents. Thus, for each of the P media contents, such as media content 2, we can select the one or more advertisements among the Q advertisements that the M objects clicked most frequently while consuming media content 2, and then map the content identifier of media content 2 to the advertisement identifiers of these one or more advertisements. Following this method, the correspondence between content identifiers and advertisement identifiers can be constructed.

[0174] In one possible implementation, the electronic device constructs the correspondence between content identifiers and advertisement identifiers through the following steps D21 and D22:

[0175] Step D21: Normalize each element in the synergy matrix to obtain the normalized synergy matrix;

[0176] Step D22: Based on the normalized collaborative matrix, construct the correspondence between content identifiers and advertising identifiers.

[0177] In this implementation, in order to easily obtain the ads that the object clicked the most when consuming media content from the collaboration matrix, each element in the collaboration matrix is ​​normalized to obtain a normalized collaboration matrix.

[0178] In some embodiments, the electronic device may employ an existing normalization method to normalize each element in the coordination matrix to obtain a normalized coordination matrix.

[0179] In some embodiments, for each element in the collaboration matrix, such as the ij-th element, the total number of consumptions of the ix-th media content corresponding to the ix-th row and the total number of clicks of the j-th advertisement corresponding to the j-th column are obtained; based on the total number of consumptions of the ix-th media content and the total number of clicks of the j-th advertisement, the normalized value of the ij-th element is determined; based on the normalized value of each element in the collaboration matrix, the normalized collaboration matrix is ​​obtained.

[0180] For example, suppose the ij-th element is Figure 5B The 12th element in the shown synergy matrix is ​​the element in the 1st row and 2nd column. For example... Figure 5B As shown, the total consumption count of the first media content C1 corresponding to the first row of the collaboration matrix is ​​6, and the total click count of the second advertisement ad2 corresponding to the second column of the collaboration matrix is ​​4. Thus, the electronic device can determine the normalized value of the 12th element based on the total consumption count of the first media content C1 (6) and the total click count of the second advertisement ad2 (4).

[0181] In one possible implementation, the electronic device multiplies the total number of times the i-th media content is consumed and the total number of times the j-th advertisement is clicked to obtain a first value; from the collaboration matrix, the number of times M objects click the j-th advertisement while consuming the i-th media content is obtained to obtain a second value; based on the first and second values, the normalized value of the ij-th element is determined.

[0182] Example 1: The ratio of the second value to the first value is determined as the normalized value of the ij-th element.

[0183] Example 2: Take the square root of the first value to obtain the third value; determine the ratio of the second value to the third value as the normalized value of the ij-th element.

[0184] In Example 2, the electronic device determines the normalized value of the ij-th element using the following formula (1):

[0185]

[0186] Where, sim ijLet N be the normalized value of the ij-th element. j (ad) represents the total number of clicks for the j-th ad, N i (tag) represents the total number of times the i-th media content is consumed, N j (ad)∩N i (tag) represents the number of times M objects click on the j-th advertisement while consuming the i-th media content; this is the second value. |N j (ad)||N i (tag)| represents the first value. It is the third value.

[0187] Referring to the above method, the electronic device can determine the normalized value of each element in the cooperation matrix, thus obtaining the normalized cooperation matrix. For example, this method is used to... Figure 5B The shown collaboration matrix is ​​normalized to obtain the normalized collaboration matrix. The value of each element in the normalized collaboration matrix can be interpreted as the relevance between the media content and the advertisement, or as the probability value of an object clicking the advertisement when consuming the media content, or as the score of the relevance between the advertisement and the media content.

[0188] In one possible implementation, the electronic device adds the total number of times the i-th media content is consumed and the total number of times the j-th advertisement is clicked to obtain a fourth value; from the collaboration matrix, the number of times M objects click the j-th advertisement while consuming the i-th media content is obtained to obtain a second value; the ratio of the second value to the fourth value is used to determine the normalized value of the ij-th element.

[0189] For example, the electronic device can be determined as the normalized value of the ij-th element by the following formula (2):

[0190]

[0191] Where, sim ij Let N be the normalized value of the ij-th element. j (ad) represents the total number of clicks for the j-th ad, N i (tag) represents the total number of times the i-th media content is consumed, N j (ad)∩N i (tag) represents the number of times M objects click on the j-th advertisement while consuming the i-th media content; this is the second value. |N j (aed)∪n i (tag)| is N j (ad) and N i(tag) The result of taking the union is the fourth value. Referring to the above method, the electronic device can determine the normalized value of each element in the coordination matrix, obtaining, as shown... Figure 5C The normalized synergy matrix is ​​shown.

[0192] It's important to note that the physical meaning of the normalized synergy matrix is ​​to calculate the correlation between advertisements and media content. This is based on the assumption that if a large number of users have simultaneously clicked on an advertisement and also viewed or searched for a particular piece of media content, then there must be some correlation between that media content and that advertisement.

[0193] After determining the normalized collaboration matrix based on the above steps, the electronic device executes step D22 to construct the correspondence between content identifiers and advertising identifiers based on the normalized collaboration matrix.

[0194] This application does not limit the specific method by which electronic devices construct the correspondence between content identifiers and advertising identifiers based on a normalized collaborative matrix.

[0195] In some embodiments, for each of the P media contents, the electronic device can, based on the normalized collaboration matrix, determine the probability value of clicking each of the Q advertisements when consuming the media content. Advertisements with a probability value of 0 are discarded, and the remaining advertisements are sorted in descending order of their probability values ​​to obtain the advertisement identifier of at least one advertisement corresponding to the content identifier of the media content.

[0196] In some embodiments, collaborative filtering outputs, such as recommendations for "beer" and "diapers," are often used. However, from an advertising recommendation perspective, such results may not be a valid strategy, and from an experience standpoint, viewing beer content followed by a diaper ad recommendation could result in a poor recommendation experience. Therefore, to ensure a strong correlation between content domain behavior and advertising, this application proposes, based on a normalized collaborative matrix, to guarantee the relevance between recommended ad text and content domain behavior text from a semantic relevance perspective. A semantic relevance model is constructed, and ad text and content domain text are vectorized using embeddings. By calculating the similarity between the ad domain and content domain embeddings, ads with similarity less than a preset value (e.g., 0.8) are filtered out. This ensures a strong correlation between recommended ads and content domain behavior text.

[0197] Specifically, for each of the P media contents, such as the k-th media content, based on the normalized collaboration matrix, T ads with the highest relevance to the k-th media content are selected from the Q ads, where k is a positive integer from 1 to P, and this relevance represents the probability of clicking the ad while consuming the k-th media content; T is a positive integer. Next, the similarity between the k-th media content and each of the T ads is determined. For example, the k-th media content and the T ads are input into a similarity calculation model, and the similarity between each ad in the k-th media content and the T ads is output. Another example is extracting the semantic feature information of the k-th media content; extracting the semantic feature information of each of the T ads; and determining the similarity between the semantic feature information of the k-th media content and the semantic feature information of each of the T ads, as the similarity between the k-th media content and each of the T ads. Then, from these T ads, S ads with the highest similarity to the k-th media content are selected, where S is a positive integer less than or equal to T. Then, the ad identifiers of the S advertisements are identified as the ad identifiers of at least one advertisement corresponding to the content identifier of the k-th media content. Following this method, the electronic device can determine the ad identifier of at least one advertisement corresponding to the content identifier of each of the P media contents, and then construct a correspondence between content identifiers and ad identifiers based on the ad identifiers of at least one advertisement corresponding to the content identifier of each of the P media contents.

[0198] For example, the correspondence between content identifiers and advertisement identifiers in this application embodiment can be represented as follows:

[0199]

[0200] In some embodiments, the correspondence between the above content identifiers and advertising identifiers can be presented in tabular form.

[0201] In this embodiment, the process of constructing the correspondence between content identifiers and ad identifiers can be completed by an offline module. In one example, after the offline module completes the construction of the correspondence between content identifiers and ad identifiers, it outputs the correspondence between content identifiers and ad identifiers to an offline table. Then, the data in the offline table is synchronously updated to the online Adtable module. The Adtable module can be an in-memory database, such as a key-value Redis.

[0202] The above describes the specific process of constructing the correspondence between content identifiers and advertising identifiers. It should be noted that the correspondence between content identifiers and advertising identifiers in this embodiment can be updated. For example, every so often, the correspondence between the constructed content identifiers and advertising identifiers is updated using the content identifiers of media content consumed by different objects in a recent period, and the advertising identifiers of advertisements clicked by different objects in a recent period, thereby ensuring the validity of the correspondence between content identifiers and advertising identifiers.

[0203] In this way, when an electronic device obtains an advertising request from a target object, after obtaining the content identifiers of N media content consumed by the target object based on the advertising request, it can query the advertising identifier corresponding to the content identifier of each of these N media content in the above-constructed correspondence between content identifiers and advertising identifiers.

[0204] For example, suppose the content identifiers of the N media content consumed by the target audience, determined based on the ad request, are content_id1 and content_id2. Assume that in the correspondence between content identifiers and ad identifiers, content_id1 corresponds to ad identifiers adgroup_id_1, adgroup_id_2, and adgroup_id_3, and content_id2 corresponds to ad identifiers adgroup_id_2 and adgroup_id_6. Thus, based on content identifiers content_id1 and content_id2, we can query the ad identifiers adgroup_id_1, adgroup_id_2, adgroup_id_3, adgroup_id_2, and adgroup_id_6 from the correspondence between content identifiers and ad identifiers. After removing the duplicate adgroup_id_2, we obtain the ad identifiers corresponding to content_id1 and content_id2 as adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6.

[0205] S103. Based on the advertising identifier corresponding to the content identifier of each of the N media contents, deliver the advertisement to the target audience.

[0206] In this embodiment of the application, the electronic device determines the advertising identifier corresponding to the content identifier of each of the N media contents based on the above steps, and then delivers the advertisement to the target object based on the advertising identifier corresponding to the content identifier of each of the N media contents.

[0207] Referring to the example above, assume that the content identifiers of N media content are content_id1 and content_id2, and the corresponding ad identifiers for content_id1 and content_id2 are adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6. Thus, the server can deliver ads to target audiences based on the ad identifiers adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6. For example, an electronic device sends an ad loading request to an advertising platform. This ad loading request includes adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6, and is used to request ad data with ad identifiers adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6. Based on this ad loading request, the advertising platform sends the ad data with ad identifiers adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6 to the electronic device. If the electronic device is a terminal device, it displays advertisements with the identifiers adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6 to the target audience in the advertising display area. If the electronic device is a server, it sends the advertising data with the identifiers adgroup_id_1, adgroup_id_2, adgroup_id_3, and adgroup_id_6 to the terminal device, and the terminal device displays these four advertisements in the advertising display area.

[0208] In some embodiments, the electronic device performs an ad recall process based on the ad identifier corresponding to the content identifier of each of the N media contents. This process may include ad removal filtering, multi-path recall merging, and ad recall truncation, to obtain the final ad to be displayed to the target audience.

[0209] The advertising delivery method provided in this application first constructs a correspondence between content identifiers and advertising identifiers. This correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each media content in multiple media content sets. This at least one advertisement is an advertisement clicked by at least one object while consuming the media content within a preset time period. Thus, when obtaining an advertising request from a target object, the content identifiers of N media content sets consumed by the target object are obtained based on the advertising request. Then, based on these N media content content identifiers, the advertising identifier corresponding to the content identifier of each of the N media content sets is queried from the correspondence between content identifiers and advertising identifiers. Finally, based on the advertising identifier corresponding to the content identifier of each of the N media content sets, advertising is delivered to the target object. Therefore, this application embodiment first constructs a correspondence between content identifiers and advertising identifiers based on the consumption behavior of different objects towards different media content and the clicking behavior of different objects towards different advertisements. This correspondence, including the advertising identifier of at least one advertisement corresponding to the content identifier of each media content in multiple media content sets, can accurately reflect the advertisements clicked by objects while consuming media content. This allows for the retrieval of advertising identifiers corresponding to the content identifiers of N media content consumed by the target audience within a specific mapping. This enables accurate prediction of advertisements that the target audience is interested in, reduces the target audience's aversion to advertisements, and improves the effectiveness of ad placement.

[0210] The overall process of the advertising delivery method according to the above application embodiment will be described below. Figure 6 and Figure 7 The advertising delivery method of the embodiments of this application will be further described.

[0211] Figure 6 This is a flowchart illustrating an advertising delivery method provided in an embodiment of this application. Figure 7 This is a schematic diagram illustrating the framework of an advertising delivery method provided in one embodiment of this application. Continuing with the example of an electronic device as the executing entity, the method of this embodiment will be described.

[0212] like Figure 6 As shown, the advertising delivery method in this application embodiment includes:

[0213] S201. Collect P media content items consumed and Q advertisements clicked by different objects within a preset time period.

[0214] Where P and Q are both positive integers greater than 1.

[0215] like Figure 7As shown, the electronic device collects the behavior of different objects in the advertising domain (e.g., clicking on advertisements) to obtain Q advertisements that different objects have clicked within a preset time period, and these Q advertisements are different advertisements. It also collects the behavior of different objects in the content domain (e.g., reading articles, watching videos, searching data, etc.) to obtain P media content that different objects have consumed within a preset time period, and these P media content are different media content.

[0216] The specific implementation process of S201 can be referred to the relevant description of step A above, and will not be repeated here.

[0217] S202. Determine the content identifier of each media content in P media content, and determine the ad identifier of each ad in Q ads.

[0218] As shown in Table 7, the electronic device determines the content identifier of each of the P media contents and places it in the content behavior table. Similarly, the electronic device determines the ad identifier of each of the Q advertisements and places it in the ad behavior table.

[0219] In this embodiment of the application, the ID of the advertisement can be directly identified as the advertisement identifier.

[0220] This application does not limit the specific method by which electronic devices determine the content identifier of media content.

[0221] In some embodiments, the method for determining the content identifier of each of the P media contents includes: for the p-th media content among the P media contents, performing content understanding on the p-th media content to obtain the content identifier of the p-th media content, where p is a positive integer from 1 to P. In one example, the electronic device first determines the content to be understood corresponding to the p-th media content. For example, if the p-th media content is obtained based on a search request, then the search keywords in the search request are determined as the content to be understood corresponding to the p-th media content. If the p-th media content is not obtained based on a search request, then the title and summary of the p-th media content are determined as the content to be understood corresponding to the p-th media content. Then, the electronic device performs content understanding on the content to be understood corresponding to the p-th media content to obtain the content identifier of the p-th media content. The process by which an electronic device understands the content to be understood corresponding to the p-th media content and obtains the content identifier of the p-th media content may include: determining the attribute values ​​of K types of attributes of the p-th media content based on the content to be understood, where K is a positive integer; and performing a hash operation on the attribute values ​​of the K types of attributes to obtain the content identifier of the p-th media content. Specifically, performing a hash operation on the attribute values ​​of the K types of attributes to obtain the content identifier of the p-th media content includes: mapping the attribute values ​​of each type of attribute in the K types of attributes to an identifier, obtaining K attribute identifier values; concatenating the K attribute identifier values ​​to obtain a concatenated attribute identifier value; and performing a hash operation on the concatenated attribute identifier value to obtain the content identifier of the p-th media content.

[0222] The specific implementation process of S202 can be referred to the relevant description of step B above, and will not be repeated here.

[0223] S203. Based on object identifiers, associate the content identifiers of P media contents with the ad identifiers of Q advertisements to obtain the content identifiers of consumed media contents and the ad identifiers of clicked advertisements for each of the M objects within a preset time period.

[0224] Where M is a positive integer greater than 1.

[0225] For example, such as Figure 7 As shown, by using the object identifier uid, the content identifiers of the media content consumed and the ad identifiers of the clicked ads of the same object within a preset time period are associated with the content identifiers of the P media contents and the ad identifiers of the Q ads, so as to obtain the content identifiers of the media content consumed and the ad identifiers of the clicked ads of each of the M objects within the preset time period.

[0226] The specific implementation process of S203 above can be referred to the relevant description of step B above, and will not be repeated here.

[0227] S204. Based on the content identifiers of consumed media content and the ad identifiers of clicked ads for each of the M objects within a preset time period, construct the correspondence between content identifiers and ad identifiers.

[0228] This application does not limit the specific method by which an electronic device constructs a correspondence between content identifiers and ad identifiers based on the content identifiers of consumed media content and the ad identifiers of clicked advertisements for each of M objects within a preset time period.

[0229] In some embodiments, the electronic device constructs a collaboration matrix based on the content identifiers of the media content consumed and the ad identifiers of the clicked ads for each of the M objects within a preset time period. This collaboration matrix represents the number of times each of the Q ads is clicked while each of the P media contents is consumed within the preset time period. Based on the collaboration matrix, a correspondence between content identifiers and ad identifiers is constructed.

[0230] In some embodiments, before constructing a collaborative matrix based on the content identifiers of consumed media content and the ad identifiers of clicked ads for each of the M objects within a preset time period, the electronic device first performs aggregation and filtering, for example, filtering out offline ads. Then, the filtered data is used to construct the correspondence between content identifiers and ad identifiers. Alternatively, before associating the content identifiers of P media contents and the ad identifiers of Q ads based on object identifiers to obtain the content identifiers of consumed media content and the ad identifiers of clicked ads for each of the M objects within a preset time period, the electronic device can filter out offline ads.

[0231] In some embodiments, the above-mentioned construction of the correspondence between content identifiers and advertising identifiers based on the collaboration matrix may include the following steps 1 and 2:

[0232] Step 1: Normalize each element in the synergy matrix to obtain the normalized synergy matrix;

[0233] Step 2: Based on the normalized collaborative matrix, construct the correspondence between content identifiers and advertising identifiers.

[0234] In some embodiments, the specific implementation process of step 1 above may include: for the ij-th element in the collaboration matrix, obtaining the total number of consumptions of the ix-th media content corresponding to the ix-th row and the total number of clicks of the j-th advertisement corresponding to the j-th column. Next, based on the total number of consumptions of the ix-th media content and the total number of clicks of the j-th advertisement, determining the normalized value of the ij-th element. For example, multiplying the total number of consumptions of the ix-th media content and the total number of clicks of the j-th advertisement yields a first value; from the collaboration matrix, obtaining the number of times M objects clicked the j-th advertisement while consuming the ix-th media content, to obtain a second value, for example, by taking the square root of the first value to obtain a third value; determining the ratio of the second value to the third value as the normalized value of the ij-th element; and based on the first and second values, determining the normalized value of the ij-th element. Finally, based on the normalized value of each element in the collaboration matrix, obtaining the normalized collaboration matrix.

[0235] In some embodiments, the specific implementation process of step 2 above may include: For the k-th media content among P media contents, firstly, based on the normalized collaboration matrix, select T advertisements from Q advertisements that have the highest correlation with the k-th media content, where k is a positive integer from 1 to P, correlation represents the probability value of clicking an advertisement while consuming the k-th media content, and T is a positive integer. Next, determine the similarity between the k-th media content and each of the T advertisements, for example, extract the semantic feature information of the k-th media content, extract the semantic feature information of each of the T advertisements, and determine the similarity between the semantic feature information of the k-th media content and the semantic feature information of each of the T advertisements, as the similarity between the k-th media content and each of the T advertisements. Then, select S advertisements from the T advertisements that have the highest similarity with the k-th media content, where S is a positive integer less than or equal to T. Finally, determine the advertisement identifiers of the S advertisements as the advertisement identifiers of at least one advertisement corresponding to the content identifier of the k-th media content. Finally, based on the content identifier of each media content in the P media content and the advertisement identifier of at least one advertisement, the correspondence between content identifiers and advertisement identifiers is constructed.

[0236] In some embodiments, the method by which the electronic device determines the advertising semantic feature information of the aforementioned T advertisements can be as follows: Figure 7 As shown, the electronic device loads the advertising data of T advertisements in batches from the advertising platform based on the advertising identifiers of the T advertisements, and then extracts the semantic features of each advertisement in the T advertisements through a semantic relevance model to obtain the semantic feature information of each advertisement in the T advertisements.

[0237] In some embodiments, the electronic device updates the correspondence between content identifiers and ad identifiers in a timely manner. In one example, ad data is loaded in batches, such as from an ad platform. The ad identifiers in the correspondence between content identifiers and ad identifiers are updated using the batch-loaded ads. For example, the ad identifiers of the batch-loaded ads are intersected with the ad identifiers in the correspondence, and the ad identifiers belonging to the batch-loaded ads in the object relationship are retained, while the ad identifiers that do not belong to the batch-loaded ads are removed.

[0238] In some embodiments, a preset period can be set to delete expired content identifiers and expired advertising identifiers in the correspondence between content identifiers and advertising identifiers.

[0239] In some embodiments, the electronic device may, at preset time intervals, reconstruct and build the latest correspondence between content identifiers and advertisement identifiers using the media content consumed and advertisements clicked by different objects within the latest time interval, according to the steps described above. In one example, the previously constructed correspondence between content identifiers and advertisement identifiers is deleted, while the latest correspondence is saved. In another example, the latest correspondence between content identifiers and advertisement identifiers is used to update the previously constructed correspondence between content identifiers and advertisement identifiers. For example, expired content identifiers of media content and expired advertisement identifiers in the previously constructed correspondence between content identifiers and advertisement identifiers are deleted, and the latest correspondence between content identifiers and advertisement identifiers is merged into the correspondence of deleted expired data, thereby obtaining the updated correspondence between content identifiers and advertisement identifiers.

[0240] In some embodiments, such as Figure 7 As shown, the electronic device outputs the constructed mapping between content identifiers and ad identifiers to an offline database (e.g., Iceberg). The data in the offline table is then synchronously updated to an online database, such as Adtable.

[0241] In some embodiments, the electronic device may also store the constructed mapping between content identifiers and ad identifiers in an offline table. When the electronic device receives an ad request from a target object, it loads the mapping between content identifiers and ad identifiers from this offline table.

[0242] The specific implementation process of S204 above can be referred to the relevant description of step C above, and will not be repeated here.

[0243] The above S201 to S204 can be understood as the process of constructing the correspondence between content identifiers and advertising identifiers.

[0244] The following steps S205 to S207 can be understood as the process of placing advertisements based on the correspondence between the constructed content identifier and the advertisement identifier.

[0245] S205. Obtain the advertising requests of the target object, and based on the advertising requests, obtain the content identifiers of the N media content consumed by the target object.

[0246] Here, the ad request is used to request an ad, and N is a positive integer.

[0247] The specific implementation process of S205 can be referred to the relevant description of S101 above, and will not be repeated here.

[0248] S206. Based on the content identifiers of N media contents, in the correspondence between content identifiers and advertising identifiers, query the advertising identifier corresponding to the content identifier of each of the N media contents.

[0249] The correspondence includes the ad identifier of at least one advertisement corresponding to the content identifier of each media content in multiple media content, and the at least one advertisement is the advertisement clicked by at least one object while consuming media content within a preset time period.

[0250] The specific implementation process of S206 can be referred to the relevant description of S102 above, and will not be repeated here.

[0251] S207. Based on the advertising identifier corresponding to the content identifier of each of the N media contents, deliver the advertisement to the target audience.

[0252] The specific implementation process of S207 can be referred to the relevant description of S103 above, and will not be repeated here.

[0253] The advertising delivery method provided in this application first collects P media content items consumed and Q advertisements clicked by different objects within a preset time period. It then determines the content identifier of each of the P media content items and the advertisement identifier of each of the Q advertisements. Next, based on the object identifiers, it associates the content identifiers of the P media content items and the advertisement identifiers of the Q advertisements to obtain the content identifiers of the media content consumed and the advertisement identifiers of the clicked advertisements for each of M objects within the preset time period. Then, based on the content identifiers of the media content consumed and the advertisement identifiers of the clicked advertisements for each of the M objects within the preset time period, it constructs a correspondence between content identifiers and advertisement identifiers. Thus, during advertising delivery, it obtains the advertisement request from the target object and, based on the advertisement request, obtains the content identifiers of the N media content items consumed by the target object. Then, based on the content identifiers of the N media content items, it queries the correspondence between the content identifiers and advertisement identifiers to find the advertisement identifier corresponding to the content identifier of each of the N media content items. Finally, based on the advertisement identifier corresponding to the content identifier of each of the N media content items, it delivers advertisements to the target object. Therefore, this embodiment first constructs a correspondence between content identifiers and ad identifiers based on the consumption behavior of different objects towards different media content and their click behavior towards different advertisements. This correspondence includes the ad identifier of at least one advertisement corresponding to the content identifier of each media content, accurately reflecting the advertisements clicked by the object when consuming media content. Thus, based on the content identifiers of N media content consumed by the target object, the corresponding ad identifiers can be retrieved from this correspondence, enabling accurate prediction of advertisements that the target object is interested in, reducing the target object's aversion to advertisements, and improving the effectiveness of ad placement.

[0254] The above text combined Figures 2 to 7 The method embodiments of this application are described in detail below, in conjunction with... Figure 8 The following describes in detail the device embodiments of this application.

[0255] Figure 8 This is a schematic block diagram of an advertising delivery device provided in an embodiment of this application.

[0256] like Figure 8 As shown, the advertising delivery device 10 includes:

[0257] The acquisition unit 11 is used to acquire the advertising request of the target object, and based on the advertising request, obtain the content identifiers of N media content consumed by the target object, wherein the advertising request is used to request an advertisement, and N is a positive integer;

[0258] The query unit 12 is used to query the advertisement identifier corresponding to the content identifier of each of the N media contents based on the content identifier of the N media contents and in the correspondence between the content identifier and the advertisement identifier. The correspondence includes the advertisement identifier of at least one advertisement corresponding to the content identifier of each of the multiple media contents. The at least one advertisement is an advertisement clicked by at least one object while consuming the media contents within a preset time period.

[0259] The processing unit 13 is used to deliver advertisements to the target object based on the advertisement identifier corresponding to the content identifier of each of the N media contents.

[0260] In some embodiments, before obtaining an ad request, the processing unit 13 is further configured to collect P media contents consumed and Q ads clicked by different objects within the preset time period, where P and Q are both positive integers greater than 1; determine the content identifier of each media content in the P media contents, and determine the ad identifier of each ad in the Q ads; associate the content identifiers of the P media contents and the ad identifiers of the Q ads based on the object identifiers to obtain the content identifiers of the media contents consumed and the ad identifiers of the ads clicked by each of M objects within the preset time period, where M is a positive integer greater than 1; and construct the correspondence between the content identifiers and the ad identifiers based on the content identifiers of the media contents consumed and the ad identifiers of the ads clicked by each of the M objects within the preset time period.

[0261] In some embodiments, the processing unit 13 is specifically used to perform content understanding on the p-th media content among the P media contents to obtain the content identifier of the p-th media content, where p is a positive integer from 1 to P.

[0262] In some embodiments, the processing unit 13 is specifically used to determine the content to be understood corresponding to the p-th media content; to perform content understanding on the content to be understood corresponding to the p-th media content, and to obtain the content identifier of the p-th media content.

[0263] In some embodiments, the processing unit 13 is specifically configured to: if the p-th media content is obtained by searching based on a search request, determine the search keywords in the search request as the content to be understood corresponding to the p-th media content; or, if the p-th media content is not obtained by searching based on a search request, determine the title and summary of the p-th media content as the content to be understood corresponding to the p-th media content.

[0264] In some embodiments, the processing unit 13 is specifically configured to determine the attribute value of the K-type attribute of the p-th media content based on the content to be understood, where K is a positive integer; and perform a hash operation on the attribute value of the K-type attribute to obtain the content identifier of the p-th media content.

[0265] In some embodiments, the processing unit 13 is specifically used to map the attribute values ​​of each of the K types of attributes to obtain K attribute identifier values; concatenate the K attribute identifier values ​​to obtain concatenated attribute identifier values; and perform a hash operation on the concatenated attribute identifier values ​​to obtain the content identifier of the p-th media content.

[0266] In some embodiments, the processing unit 13 is specifically configured to construct a collaboration matrix based on the content identifier of the media content consumed and the ad identifier of the clicked ads for each of the M objects within the preset time period. The collaboration matrix represents the number of times the M objects click each of the Q ads while consuming each of the P media contents within the preset time period. Based on the collaboration matrix, a correspondence between the content identifier and the ad identifier is constructed.

[0267] In some embodiments, the processing unit 13 is specifically configured to determine, based on the content identifier of the media content consumed and the ad identifier of the clicked ads by each of the M objects within the preset time period, the number of times each of the P media contents is consumed and each of the Q ads is clicked by each of the M objects within the preset time period; and to construct the collaboration matrix based on the number of times each of the P media contents is consumed and each of the Q ads is clicked by each of the M objects within the preset time period, as well as the content identifier of the P media contents and the ad identifier of the Q ads, wherein the ij-th element in the collaboration matrix represents the number of times the M objects consume the ith media content and click the j-th ad, where i is a positive integer less than or equal to P and j is a positive integer less than or equal to Q.

[0268] In some embodiments, the processing unit 13 is specifically used to normalize each element in the collaboration matrix to obtain a normalized collaboration matrix; and to construct the correspondence between the content identifier and the advertisement identifier based on the normalized collaboration matrix.

[0269] In some embodiments, if the collaboration matrix is ​​a P-row, Q-column matrix, the processing unit 13 is specifically configured to, for the ij-th element in the collaboration matrix, obtain the total consumption count of the i-th media content corresponding to the i-th row and the total click count of the j-th advertisement corresponding to the j-th column; determine the normalized value of the ij-th element based on the total consumption count of the i-th media content and the total click count of the j-th advertisement; and obtain the normalized collaboration matrix based on the normalized value of each element in the collaboration matrix.

[0270] In some embodiments, the processing unit 13 is specifically configured to multiply the total number of consumptions of the i-th media content and the total number of clicks of the j-th advertisement to obtain a first value; obtain from the collaboration matrix the number of times the M objects clicked the j-th advertisement while consuming the i-th media content to obtain a second value; and determine the normalized value of the ij-th element based on the first value and the second value.

[0271] In some embodiments, the processing unit 13 is specifically used to take the square root of the first value to obtain a third value; and to determine the ratio of the second value to the third value as the normalized value of the ij-th element.

[0272] In some embodiments, the processing unit 13 is specifically configured to, for the k-th media content among P media contents, based on the normalized collaboration matrix, select T advertisements from the Q advertisements that have the highest correlation with the k-th media content, where k is a positive integer from 1 to P, the correlation represents the probability value of clicking the advertisement while consuming the k-th media content, and T is a positive integer; determine the similarity between the k-th media content and each of the T advertisements; select S advertisements from the T advertisements that have the highest similarity with the k-th media content, where S is a positive integer less than or equal to T; determine the advertisement identifiers of the S advertisements as the advertisement identifiers of at least one advertisement corresponding to the content identifier of the k-th media content; and construct the correspondence between the content identifier and the advertisement identifier based on the advertisement identifiers of at least one advertisement corresponding to the content identifier of each media content among the P media contents.

[0273] In some embodiments, the processing unit 13 is specifically used to extract the semantic feature information of the k-th media content; extract the semantic feature information of each of the T advertisements; and determine the similarity between the semantic feature information of the k-th media content and the semantic feature information of each of the T advertisements, as the similarity between the k-th media content and each of the T advertisements.

[0274] It should be understood that the device embodiments and method embodiments can correspond to each other, and similar descriptions can be referred to the method embodiments. To avoid repetition, further details will not be provided here. Specifically, Figure 8 The apparatus shown can perform the embodiments of the above-described method, and the foregoing and other operations and / or functions of each module in the apparatus are respectively for implementing the embodiments of the above-described model training method. For the sake of brevity, they will not be described in detail here.

[0275] The apparatus of this application embodiment has been described above from the perspective of functional modules in conjunction with the accompanying drawings. It should be understood that this functional module can be implemented in hardware, in software instructions, or in a combination of hardware and software modules. Specifically, the steps of the method embodiments in this application can be completed by integrated logic circuits in the processor's hardware and / or by software instructions. The steps of the method disclosed in this application embodiment can be directly embodied as being executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. Optionally, the software module can reside in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps in the above method embodiments.

[0276] Figure 9 This is a schematic block diagram of an electronic device provided in the embodiments of this application. The electronic device can be the terminal device or server described above, used to execute the method embodiments described above.

[0277] like Figure 9 As shown, the electronic device 40 may include:

[0278] The system includes a memory 41 and a processor 42. The memory 41 stores a computer program 43 and transfers the program code 43 to the processor 42. In other words, the processor 42 can retrieve and run the computer program 43 from the memory 41 to implement the methods described in the embodiments of this application.

[0279] For example, the processor 42 can be used to execute the steps in the above method according to the instructions in the computer program 43.

[0280] In some embodiments of this application, the processor 42 may include, but is not limited to:

[0281] General-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0282] In some embodiments of this application, the memory 41 includes, but is not limited to:

[0283] Volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0284] In some embodiments of this application, the computer program 43 may be divided into one or more modules, which are stored in the memory 41 and executed by the processor 42 to complete the page recording method provided in this application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 43 in the electronic device.

[0285] like Figure 9 As shown, the electronic device 40 may further include:

[0286] Transceiver 44, which can be connected to processor 42 or memory 41.

[0287] The processor 42 can control the transceiver 44 to communicate with other devices; specifically, it can send information or data to other devices or receive information or data sent by other devices. The transceiver 44 may include a transmitter and a receiver. The transceiver 44 may further include antennas, and the number of antennas may be one or more.

[0288] It should be understood that the various components in the electronic device 40 are connected through a bus system, which includes a data bus, a power bus, a control bus, and a status signal bus.

[0289] According to one aspect of this application, a computer storage medium is provided that stores a computer program thereon, which, when executed by a computer, enables the computer to perform the methods of the above-described method embodiments. Alternatively, embodiments of this application also provide a computer program product containing instructions that, when executed by a computer, cause the computer to perform the methods of the above-described method embodiments.

[0290] According to another aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method described in the above-described method embodiments.

[0291] In other words, when implemented using software, it can be implemented wholly or partially in the form of a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0292] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0293] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0294] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. For example, the functional modules in the various embodiments of this application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0295] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An advertising placement method, characterized in that, include: Obtain the advertising request of the target object, and based on the advertising request, obtain the content identifiers of N media content consumed by the target object, wherein the advertising request is used to request an advertisement, and N is a positive integer; Based on the content identifiers of the N media contents, in the correspondence between content identifiers and advertising identifiers, query the advertising identifier corresponding to the content identifier of each of the N media contents. The correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each of the multiple media contents. The at least one advertisement is an advertisement clicked by at least one object while consuming the media contents within a preset time period. Based on the advertising identifier corresponding to the content identifier of each of the N media contents, advertisements are delivered to the target audience.

2. The method according to claim 1, characterized in that, Before obtaining the advertisement request, the method further includes: Collect P media content items consumed and Q advertisements clicked by different objects within the preset time period, where P and Q are both positive integers greater than 1; Determine the content identifier of each of the P media contents, and determine the advertisement identifier of each of the Q advertisements; Based on the object identifier, the content identifiers of the P media contents and the ad identifiers of the Q advertisements are associated to obtain the content identifiers of the consumed media contents and the ad identifiers of the clicked advertisements for each of the M objects within the preset time period, where M is a positive integer greater than 1. Based on the content identifiers of consumed media content and the ad identifiers of clicked advertisements for each of the M objects within the preset time period, a correspondence between the content identifiers and the ad identifiers is constructed.

3. The method according to claim 2, characterized in that, Determining the content identifier of each of the P media contents includes: For the p-th media content among the P media content, content understanding is performed on the p-th media content to obtain the content identifier of the p-th media content, where p is a positive integer from 1 to P.

4. The method according to claim 3, characterized in that, The step of performing content understanding on the p-th media content to obtain the content identifier of the p-th media content includes: Determine the content to be understood corresponding to the p-th media content; The content to be understood corresponding to the p-th media content is understood to obtain the content identifier of the p-th media content.

5. The method according to claim 4, characterized in that, Determining the content to be understood corresponding to the p-th media content includes: If the p-th media content is obtained based on a search request, then the search keywords in the search request are determined as the content to be understood corresponding to the p-th media content; or, If the p-th media content is not obtained through a search request, then the title and description of the p-th media content are determined as the content to be understood corresponding to the p-th media content.

6. The method according to claim 4, characterized in that, The step of performing content understanding on the content to be understood corresponding to the p-th media content to obtain the content identifier of the p-th media content includes: Based on the content to be understood, determine the attribute value of the K-type attribute of the p-th media content, where K is a positive integer; The attribute values ​​of each of the K types of attributes are mapped to identifiers to obtain K attribute identifier values; The K attribute identifier values ​​are concatenated to obtain the concatenated attribute identifier value; A hash operation is performed on the concatenated attribute identifier value to obtain the content identifier of the p-th media content.

7. The method according to claim 2, characterized in that, The step of constructing a correspondence between content identifiers and ad identifiers based on the content identifiers of consumed media content and the ad identifiers of clicked ads for each of the M objects within the preset time period includes: Based on the content identifier of the media content consumed and the ad identifier of the clicked ad for each of the M objects within the preset time period, a collaboration matrix is ​​constructed. The collaboration matrix represents the number of times the M objects click each of the Q ads while consuming each of the P media contents within the preset time period. Based on the aforementioned collaboration matrix, a correspondence between the content identifier and the advertising identifier is constructed.

8. The method according to claim 7, characterized in that, The step of constructing a collaborative matrix based on the content identifiers of consumed media content and the ad identifiers of clicked ads for each of the M objects within the preset time period includes: Based on the content identifier of the media content consumed and the ad identifier of the ads clicked by each of the M objects within the preset time period, determine the number of times each of the Q ads is clicked by each of the P media contents consumed by each of the M objects within the preset time period. Based on the number of times each of the P media contents consumed by the M objects during the preset time period, and the content identifiers of the P media contents and the ad identifiers of the Q ads, the collaboration matrix is ​​constructed. The ij-th element in the collaboration matrix represents the number of times the M objects consume the ith media content and click the j-th ad, where i is a positive integer less than or equal to P, and j is a positive integer less than or equal to Q.

9. The method according to claim 8, characterized in that, The step of constructing the correspondence between content identifiers and advertisement identifiers based on the collaboration matrix includes: Normalize each element in the synergy matrix to obtain the normalized synergy matrix; Based on the normalized collaborative matrix, the correspondence between the content identifier and the advertisement identifier is constructed.

10. The method according to claim 9, characterized in that, If the cooperative matrix is ​​a P-row, Q-column matrix, the normalization process for each element of the cooperative matrix to obtain a normalized cooperative matrix includes: For the ij-th element in the collaboration matrix, obtain the total number of consumptions of the ij-th media content corresponding to the ij-th row in the collaboration matrix, and the total number of clicks of the j-th advertisement corresponding to the j-th column; Based on the total number of consumptions of the i-th media content and the total number of clicks on the j-th advertisement, determine the normalized value of the ij-th element; The normalized collaborative matrix is ​​obtained based on the normalized value of each element in the collaborative matrix.

11. The method according to claim 10, characterized in that, The process of determining the normalized value of the ij-th element based on the total number of consumptions of the i-th media content and the total number of clicks on the j-th advertisement includes: The first value is obtained by multiplying the total number of consumptions of the i-th media content and the total number of clicks of the j-th advertisement; From the collaboration matrix, obtain the number of times the M objects clicked the j-th advertisement while consuming the i-th media content, to obtain a second value; Based on the first value and the second value, the normalized value of the ij-th element is determined.

12. The method according to claim 9, characterized in that, The step of constructing the correspondence between content identifiers and advertisement identifiers based on the normalized collaborative matrix includes: For the k-th media content among P media content, based on the normalized collaboration matrix, select T ads from the Q ads that have the greatest correlation with the k-th media content, where k is a positive integer from 1 to P, the correlation represents the probability value of clicking the ad while consuming the k-th media content, and T is a positive integer; Determine the similarity between the k-th media content and each of the T advertisements; From the T advertisements, select the S advertisements that have the highest similarity to the content of the kth media, where S is a positive integer less than or equal to T; The advertising identifiers of the S advertisements are determined as the advertising identifiers of at least one advertisement corresponding to the content identifier of the kth media content; Based on the content identifier of each of the P media contents and the advertisement identifier of at least one advertisement corresponding to the content identifier, the correspondence between the content identifier and the advertisement identifier is constructed.

13. An advertising delivery device, characterized in that, include: The acquisition unit is used to acquire the advertising request of the target object, and based on the advertising request, obtain the content identifiers of N media content consumed by the target object, wherein the advertising request is used to request an advertisement, and N is a positive integer; The query unit is used to query the advertising identifier corresponding to the content identifier of each of the N media contents based on the content identifier of the N media contents and in the correspondence between content identifier and advertising identifier. The correspondence includes the advertising identifier of at least one advertisement corresponding to the content identifier of each of the multiple media contents. The at least one advertisement is an advertisement clicked by at least one object while consuming the media contents within a preset time period. The processing unit is used to deliver advertisements to the target object based on the advertisement identifier corresponding to the content identifier of each of the N media contents.

14. An electronic device, comprising a processor and a memory; The memory is used to store computer programs; The processor is configured to execute the computer program to implement the method as described in any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that, Used to store computer programs; The computer program causes the computer to perform the method as described in any one of claims 1 to 12.

Images