A method and device for processing popularization content, a computer device and a storage medium

Abstract

Embodiments of the present application disclose a promotion content processing method and device, computer equipment and a storage medium. A promotion content request of an object can be acquired. A promotion content set is determined according to the promotion content request, wherein the promotion content set includes at least one promotion content. If the behavior characteristics of the object with respect to the promotion content meet a preset negative condition, the historical behavior characteristic data of the object with respect to the promotion content is detected. When the object has historical behavior characteristic data, content recall is performed on the promotion content set based on the historical behavior characteristic data to obtain a target promotion content set of the object. When the object does not have historical behavior characteristic data, content sorting processing is performed on the promotion content set to recall the promotion content of the object to obtain the target promotion content set of the object. The target promotion content of the object is determined from the target promotion content set, and the target promotion content is sent to the object. The scheme can improve the promotion content processing efficiency for the object.

Description

Technical Field

[0001] This application relates to the field of computer technology, and specifically to a method, apparatus, computer device, and storage medium for processing promotional content. Background Technology

[0002] Promotion can be used to expand the reach of something; for example, marketing, publicity, and advertising can all be forms of promotion. Correspondingly, promotional content can be anything needed for promotion, such as text content, images, videos, audio content, and links.

[0003] In the process of researching and practicing related technologies, the inventors of this application discovered that when promoting content to a target audience, some audiences exhibit negative content interaction behavior towards the promoted content, resulting in low accuracy and efficiency of existing promotional content processing methods for that audience. Therefore, the methods for processing promotional content need to be improved. Summary of the Invention

[0004] This application provides a method, apparatus, computer device, and storage medium for processing promotional content, thereby improving the efficiency of processing promotional content targeting specific audiences.

[0005] This application provides a method for processing promotional content, including: Request the promotional content of the object; A set of promotional content is determined based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content; If the object's behavioral characteristics toward the promotional content meet the preset negative behavioral conditions, then the object's historical behavioral characteristic data toward the promotional content will be detected. When the object has the historical behavior feature data, content recall is performed on the set of promotional content based on the historical behavior feature data to obtain the target set of promotional content for the object. When the object does not have the historical behavior feature data, the promotion content set is sorted to recall the object's promotion content and obtain the object's target promotion content set. The target promotion content for the object is determined from the set of target promotion content, and the target promotion content is sent to the object.

[0006] Accordingly, embodiments of this application also provide a processing apparatus for promotional content, including: The retrieval unit is used to retrieve the promotional content request for an object. A set determination unit is used to determine a set of promotional content based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content; The detection unit is used to detect the historical behavioral feature data of the object in relation to the promotional content if the object's behavioral features in relation to the promotional content meet the preset negative behavioral conditions. The first recall unit is used to recall content based on the historical behavior feature data of the object when the object has the historical behavior feature data, so as to obtain the target promotion content set of the object. The second recall unit is used to sort the promotional content set when the object does not have the historical behavior feature data, so as to recall the object's promotional content and obtain the object's target promotional content set. The content determination unit is used to determine the target promotion content of the object from the target promotion content set and send the target promotion content to the object.

[0007] In one embodiment, the detection unit includes: The first acquisition subunit is used to acquire the object's behavior information in response to the promotional content; The first analysis subunit is used to perform statistical analysis on the behavioral information to obtain the behavioral characteristics of the object in response to the promotional content. The detection subunit is used to detect the historical behavioral feature data of the object in relation to the promotional content if the behavioral features meet the preset negative behavioral conditions.

[0008] In one embodiment, the first acquisition subunit is configured to: Adjust the content promotion evaluation threshold in the content promotion system; based on the adjusted content promotion evaluation threshold, send promotional content to the target; obtain the target's behavior information regarding the promotional content within a preset time interval.

[0009] In one embodiment, the first recall unit includes: The preference determination subunit is used to determine the object's preference information for the promoted content based on the historical behavioral feature data. The first recall subunit is used to recall content based on the preference information for the set of promotional content to obtain the target set of promotional content for the object.

[0010] In one embodiment, the first recall subunit is configured to: Based on the preference information, the object group to which the object belongs is determined; promotional content associated with the object group is selected from the promotional content set; the selected promotional content is sorted to obtain sorted promotional content; based on the sorting result, content retrieval is performed on the sorted promotional content to obtain the target promotional content set for the object.

[0011] In one embodiment, the first recall subunit is configured to: Obtain the content description information of the promotional content; based on the preference information and the content description information, perform content preference matching between the object and the promotional content; based on the matching result, perform content retrieval for the set of promotional content to obtain the target set of promotional content for the object.

[0012] In one embodiment, the first recall unit includes: The model defines a sub-unit used to determine the original promotion evaluation model for the promoted content; The feature extraction subunit is used to extract features from the historical behavior feature data to obtain the content interaction features of the object in response to the promotional content; The feature addition subunit is used to add the content interaction features to the original promotion evaluation model, so that the promotion evaluation model after addition can determine the content promotion evaluation result of the promotion content corresponding to the object by referring to the content interaction features; The second recall subunit is used to recall the promotion content set based on the content promotion evaluation results, so as to obtain the target promotion content set of the object.

[0013] In one embodiment, the second recall unit includes: The information acquisition subunit is used to acquire promotion statistics of the promotional content and the object's preference information for the promotional content; The parameter determination subunit is used to determine content promotion evaluation parameters for the promoted content based on the promotion statistics and the preference information. The first sorting subunit is used to sort the promotion content set based on the content promotion evaluation parameters in order to recall the promotion content of the object and obtain the target promotion content set of the object.

[0014] In one embodiment, the first sorting subunit is configured to: Determine the original promotion evaluation model for the promoted content; add the content promotion evaluation parameters to the original promotion evaluation model so that the added promotion evaluation model determines the content promotion evaluation result of the promoted content corresponding to the object based on the content promotion evaluation parameters; perform content sorting processing on the set of promoted content based on the content promotion evaluation result; recall the promoted content of the object based on the sorting result to obtain the target promotion content set of the object.

[0015] In one embodiment, the second recall unit includes: The attribute acquisition unit is used to acquire the content attribute information of the promotional content and the object attribute information of the object; The second sorting subunit is used to sort the promotional content set based on the content attribute information and the object attribute information in order to recall the promotional content of the object and obtain the target promotional content set of the object.

[0016] In one embodiment, the second sorting subunit is used for: Based on the content attribute information and the object attribute information, select promotional content to be sorted from the promotional content set; sort the promotional content to be sorted to obtain sorted promotional content; based on the sorting result, recall the promotional content of the object from the sorted promotional content to obtain the target promotional content set of the object.

[0017] In one embodiment, the apparatus for processing the promotional content further includes: A feature determination unit is used to determine the target behavioral features of the object in relation to the target promotional content; The type determination unit is used to determine the object type of the object for the promotional content based on the target behavior characteristics and the preset negative behavior conditions.

[0018] Accordingly, this application also provides a storage medium storing a computer program thereon, wherein when the computer program is executed by a processor, it implements the steps of the promotional content processing method shown in the embodiments of this application.

[0019] Accordingly, this application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the processing method for promotional content as shown in the embodiments of this application.

[0020] This application embodiment can obtain a target's promotional content request; determine a set of promotional content based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content; if the target's behavioral characteristics toward the promotional content meet preset negative behavioral conditions, detect the target's historical behavioral characteristic data toward the promotional content; when the target exists, perform content recall on the set of promotional content based on the historical behavioral characteristic data to obtain the target set of promotional content for the target; when the target does not exist, perform content sorting on the set of promotional content to recall the target's promotional content to obtain the target set of promotional content for the target; determine the target promotional content for the target from the target set of promotional content, and send the target promotional content to the target.

[0021] This solution targets objects whose behavioral characteristics in response to promoted content meet preset negative behavioral conditions. After receiving a request for promoted content from a target object, it determines how to retrieve promoted content from the target object's content set based on whether the target object has historical behavioral data related to the promoted content. This allows for precise content promotion to the target object. Specifically, if the target object has historical behavioral data related to the promoted content, the solution can retrieve content based on this data. This allows for accurate prediction of the promoted content that the target object might be interested in by referring to the target object's content interaction history. If the target object does not have historical behavioral data related to the promoted content, the solution can retrieve corresponding promoted content from the promoted content set through sorting exploration, increasing the probability of predicting the promoted content that the target object might be interested in. Therefore, this solution improves the accuracy and efficiency of content promotion to the target object by increasing the processing efficiency of promoted content. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application, 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram illustrating a scenario of the method for processing promotional content provided in an embodiment of this application; Figure 2 This is a flowchart of the method for processing promotional content provided in the embodiments of this application; Figure 3This is another flowchart illustrating the method for processing promotional content provided in the embodiments of this application; Figure 4 This is another flowchart illustrating the method for processing promotional content provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of the promotional content processing device provided in the embodiments of this application; Figure 6 This is another schematic diagram of the structure of the promotional content processing device provided in the embodiments of this application; Figure 7 This is another schematic diagram of the structure of the promotional content processing device provided in the embodiments of this application; Figure 8 This is another schematic diagram of the structure of the promotional content processing device provided in the embodiments of this application; Figure 9 This is another schematic diagram of the structure of the promotional content processing device provided in the embodiments of this application; Figure 10 This is another schematic diagram of the structure of the promotional content processing device provided in the embodiments of this application; Figure 11 This is another schematic diagram of the structure of the promotional content processing device provided in the embodiments of this application; Figure 12 This is a schematic diagram of the structure of the computer device provided in the embodiments of this application. Detailed Implementation

[0024] 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 skilled in the art without creative effort are within the scope of protection of this application.

[0025] This application provides a method, apparatus, computer device, and storage medium for processing promotional content. Specifically, this application provides an apparatus for processing promotional content suitable for computer devices. The computer device can be a terminal or a server, and the terminal can be a mobile phone, tablet computer, laptop computer, or similar device. The server can be a single server or a server cluster consisting of multiple servers.

[0026] This application embodiment will take the example of a method for processing promotional content jointly executed by a server and a terminal to introduce the method for processing promotional content.

[0027] refer to Figure 1An object can send a promotional content request to the server 10 via the terminal 20. In response, the server 10 can obtain the object's promotional content request and determine a set of promotional content based on the request. This set of promotional content may include at least one piece of promotional content. If the object's behavioral characteristics regarding the promotional content meet preset negative behavioral conditions, the server 10 can detect the object's historical behavioral characteristic data regarding the promotional content and, based on whether the object possesses such historical behavioral characteristic data, perform corresponding content recall to obtain the object's target set of promotional content.

[0028] Specifically, when the object has the historical behavior feature data, the server 10 can recall the promotional content set based on the historical behavior feature data to obtain the target promotional content set for the object; when the object does not have the historical behavior feature data, the server 10 can sort the promotional content set to recall the promotional content for the object to obtain the target promotional content set for the object.

[0029] Furthermore, server 10 can determine the target promotion content for the object from the target promotion content set and send the target promotion content to the object. For example, the target promotion content can be sent to terminal 20 so that the object can obtain the target promotion content through terminal 20.

[0030] The following sections provide detailed descriptions of each example. It should be noted that the order in which the embodiments are described is not intended to limit the preferred order of the embodiments.

[0031] This application provides a method for processing promotional content. This method can be executed by a terminal or a server, or by both. This application example illustrates the method by having the server execute the promotional content processing method; specifically, it is executed by a promotional content processing device integrated into the server. Figure 2 The specific process for processing this promotional content can be as follows: 101. Obtain the target's promotional content request.

[0032] Promotion can be used to expand the scope of influence of something; for example, marketing, publicity, and advertising can all be forms of promotion.

[0033] Correspondingly, promotional content can be anything needed for promotion. For example, promotional content can include several different forms of data; that is, promotional content can exist in multiple forms, such as text content, image content, video content, audio content, and link content. Furthermore, promotional content can be obtained by combining several types of content, allowing the resulting promotional content to include several different forms of data. For example, taking an advertisement as an example, the advertisement content can include multiple elements such as advertisement description text, images, videos, audio, and interactive settings; and so on.

[0034] In the advertising example, the promoted content can specifically be in-feed ads. In-feed ads are ads that appear in the feeds of social media users' friends or in the content streams of news and audiovisual media. In-feed ads can take the form of images, text and images, and videos, and are characterized by algorithmic recommendations, a native user experience, and the ability to target ads through tags, allowing users to choose whether to push exposure, landing pages, or app downloads based on their needs. In-feed ads can be used in various media, including news, social media, video, and search engines.

[0035] As examples, news media can include news apps. For news media, advertisements can be inserted between messages without affecting the user experience. Social media can include social apps. For social media, advertisements can be inserted between user status feeds, and users can interact with the ads by liking, sharing, and commenting. Video media can include video apps. For video media, ads can be inserted between videos, and users will see the ads when searching for videos. Search media can include search apps. For search media, feed ads rely on search engines; when users search for keywords through a search engine, several relevant ad recommendation links will be automatically matched in the search results.

[0036] The object refers to the entity that uses the application or service. For example, the object may include the customers who use the application or service, such as individual users or organizations.

[0037] Among them, the promotion content request is a request to obtain relevant information data of promotion content. For example, an object can send a promotion content request to the server through the terminal, so that the server can respond to the promotion content request and send the corresponding promotion content to the terminal, thereby realizing content promotion to the object.

[0038] As an example, if the content being promoted is an advertisement, such as a news feed ad, and the target audience is a user, the server can obtain the user's ad request so that the server can respond to the ad request and send the corresponding ad to the user.

[0039] There are several ways for a server to obtain promotional content requests for an object. For example, it can obtain it through a terminal; or it can obtain it through other servers; and so on.

[0040] 102. Determine the set of promotional content based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content.

[0041] The promotion content set is a set that includes at least one promotion content. For example, if the promotion content is an advertisement, then the promotion content set can be an advertisement set, which may include at least one advertisement.

[0042] As an example, the collection of promotional content can include a collection of historical promotional content and a collection of content to be promoted.

[0043] Specifically, the content to be promoted refers to the promotional content to be promoted to the target audience. For example, the content to be promoted can be an advertisement to be promoted to users. Correspondingly, the set of content to be promoted is a collection of content to be promoted, which may include at least one piece of content to be promoted. For example, the set of content to be promoted can be a set of advertisements to be promoted, which may include at least one advertisement to be promoted.

[0044] Historical promotional content refers to promotional content that has already been promoted to the target audience. For example, historical promotional content can be advertisements that have been promoted to users. Correspondingly, the historical promotional content set is a collection of historical promotional content, which may include at least one historical promotional content. For example, the historical promotional content set may be a collection of historical promotional advertisements, which may include at least one historical promotional advertisement.

[0045] There are several ways to determine the set of promotional content based on a promotional content request. For example, after receiving the promotional content request, one can request the set of promotional content from a content promotion system. The content promotion system is a system used to promote content to an audience. This system can run as a client on a terminal or server, and it can store promotional content, such as historical promotional content and content to be promoted. As an example, the promotional content can be an advertisement, and the content promotion system can be an advertising platform running on a server, capable of promoting advertisements.

[0046] For example, different promotional content sets can be set for different object types. As an example, the object identifier of an object can be carried in the promotional content request of the object. The server can determine the object type of the object based on the object identifier and determine the promotional content set corresponding to the object type as the promotional content set corresponding to the promotional content request.

[0047] For example, different sets of promotional content can be set for different times. As an example, different sets of promotional content can be set for different time periods. After obtaining the promotional content request of an object, the set of promotional content corresponding to the current time period can be determined and set as the set of promotional content corresponding to the promotional content request.

[0048] For example, based on different request keywords, the content promotion system can select promotional content related to those keywords to generate a set of promotional content. As an example, an object's promotional content request can carry its request keywords, such as search keywords. The content promotion system can then select search content related to those search keywords from the stored promotional content to generate a set of promotional content corresponding to that request; and so on.

[0049] 103. If the object's behavioral characteristics toward the promotional content meet the preset negative behavioral conditions, then the historical behavioral characteristic data of the object toward the promotional content will be detected.

[0050] Among them, the behavioral characteristics of the object in relation to the promoted content are the relevant features describing the object's content interaction behavior in relation to the promoted content. For example, the content interaction behavior in relation to the promoted content may include specific touch operations, such as single click, long press, double click, and swipe, etc., and may also be triggered by voice; optionally, the content interaction behavior may also be a combination of a series of operations.

[0051] In one embodiment, the promotional content can be an advertisement, and the target can be a user. The behavioral characteristics of the target in relation to the promotional content can be the relevant characteristics of the user's advertising interaction behavior. For example, advertising interaction behavior can include a user's ad click behavior, ad comment behavior, ad favorite behavior, ad product purchase behavior, etc.

[0052] The preset negative behavior condition is used to determine whether an object has negative content interaction behavior towards the promoted content. Negative describes content interaction behavior that is the opposite of active behavior. Negative content interaction behavior can refer to an object's passive reaction to the promoted content, such as not clicking on the exposed promoted content, closing the exposed promoted content within a predetermined time threshold, or blocking the promoted content. Therefore, the preset negative behavior condition can be set based on the object's content interaction behavior towards the promoted content.

[0053] In one embodiment, the promotional content can be an advertisement, the target audience can be users, and the preset negative behavior conditions can be set based on users' ad click behavior, ad comment behavior, ad collection behavior, ad product purchase behavior, etc. For example, the preset negative behavior conditions can be: the ad is exposed more than 60 times in the same media within a predetermined time period (e.g., the most recent 30 days), and there is no ad click behavior.

[0054] Among them, historical behavioral feature data refers to feature data describing historical behavior. Here, historical behavior refers to the user's historical interest in the promoted content. For example, historical behavior can include a user's historical behavior on a news application, which can describe the user's interest in the advertisement. This historical behavior may include historical browsing behavior, historical liking behavior, historical commenting behavior, historical purchasing behavior, etc. Corresponding historical behavioral feature data can include feature data on the time of the behavior, such as the time of browsing behavior; feature data on the duration of the behavior, such as the duration of commenting behavior; feature data describing the type of content the behavior pertains to, such as the type of news the user browsed; and so on.

[0055] It is worth noting that, in this application, the historical behavioral characteristic data of the object for the promotional content may include the historical behavioral characteristic data of the object in different applications, for example, it may include the historical behavioral characteristic data of users in different applications.

[0056] In one embodiment, taking the promoted content as an advertisement and the target as a user as an example, users whose behavioral characteristics towards the advertisement meet preset negative behavioral conditions can be called silent users. For example, a silent user can be a user whose advertisement has been exposed more than 60 times in the same media in the past 30 days without clicking on the advertisement. In this embodiment, the user terminal can send an advertisement request message to the content promotion system in the server. The advertisement request message may include a user identifier. When the content promotion system receives the advertisement request message, it can determine the behavioral information of the user corresponding to the user identifier based on the user identifier in the advertisement request message, and determine the user's behavioral characteristics towards the advertisement based on the behavioral information. Then, it can determine whether the user's behavioral characteristics towards the advertisement meet the preset negative behavioral conditions. When the user's behavioral characteristics towards the advertisement meet the preset negative behavioral conditions, it can be determined that the user is a silent user, and the user's historical behavioral characteristic data towards the advertisement can be detected.

[0057] Furthermore, in some embodiments, at least one target object can be pre-identified and stored in the content promotion system. When the content promotion system receives a promotional content request from a terminal, it determines whether the user corresponding to the promotional content request is a target object among the at least one target object. If the user corresponding to the promotional content request is a target object among the at least one target object, then the user corresponding to the promotional content request is identified as the target object. In this way, the speed at which the content promotion system identifies target objects can be improved, thereby increasing the speed of content promotion.

[0058] In other embodiments, the content promotion system may also determine the user corresponding to the promotion content request after receiving it from the terminal, determine the user's behavior information regarding the advertisement, and determine whether the user's behavior characteristics regarding the advertisement meet preset negative behavior conditions based on the behavior information. When the user's behavior characteristics regarding the advertisement meet the preset negative behavior conditions, the user is determined to be a target, and the system further detects the user's historical behavior characteristic data regarding the advertisement. This approach can improve the real-time performance and accuracy of the content promotion system in determining the target. Specifically, the step "if the target's behavior characteristics regarding the promotion content meet the preset negative behavior conditions, then detect the target's historical behavior characteristic data regarding the promotion content" may include: Obtain information about the target's behavior in response to the promoted content; Statistical analysis of behavioral information yields the behavioral characteristics of the target audience in response to the promotional content; If the behavioral characteristics meet the preset negative behavioral conditions, then the historical behavioral characteristic data of the target object in relation to the promoted content will be detected.

[0059] The "object's behavioral information regarding the promoted content" refers to information related to the object's content interaction behavior with the promoted content. For example, it could be statistical information describing the object's historical content interaction behavior with the promoted content. Content interaction behavior can include specific touch operations on the promoted content, such as single-click, long-press, double-click, and swipe operations, etc. Content interaction behavior can also be triggered by voice; optionally, content interaction behavior can also be a combination of a series of operations.

[0060] As an example, the promotional content can be an advertisement, and the target can be a user. The user's behavioral information in relation to the promotional content can be statistical information on the user's historical advertising interaction behavior. For example, the user's historical advertising interaction behavior can include the user's past advertising click behavior, advertising comment behavior, advertising collection behavior, advertising product purchase behavior, etc. Correspondingly, the user's behavioral information can include statistical information on the user's historical advertising interaction behavior.

[0061] There are several ways to obtain information about an object's behavior in relation to the promoted content. For example, the object can report its behavior information to the server so that the server can obtain the information. Alternatively, the content promotion system can collect information about the object's behavior in relation to the promoted content, and report it to the server, provided it has the data collection permission and complies with relevant regulations. Another example is that one or more servers hosting the content promotion system can store the objects targeted by the promoted content and their behavior information, allowing the system to use the targeted object as the target and retrieve its behavior information from the server's storage system.

[0062] In this application, there are various ways to perform statistical analysis on behavioral information. For example, the behavioral information can be statistically analyzed based on the definition of behavioral characteristics. For instance, behavioral characteristics may include statistical features such as the total number, maximum value, and average value of behaviors targeting promotional content within a preset time interval. Therefore, the behavioral information can be statistically analyzed based on the specific definition of behavioral characteristics to obtain the corresponding behavioral characteristics.

[0063] In one embodiment, the content promotion system can be an advertising platform A, the promoted content can be advertisements, the target audience can be users, and the preset negative behavior condition can be: the number of ad impressions in the same application exceeds 60 times in the past 30 days, and there is no ad click behavior. Therefore, the server can obtain the number of ad impressions of a user in the same application, as well as the user's ad click behavior information, and based on the ad click behavior information, count the number of ad click behaviors of the user in the past 30 days, and use the statistical results as the user's ad-related behavior characteristics.

[0064] Furthermore, if an object's behavioral characteristics meet preset negative behavioral conditions, then the object's historical behavioral characteristics data regarding the promoted content can be detected. For example, if user B has had more than 60 ad impressions in application A within the last 30 days without clicking on any ads, then user B can be identified as a dormant user of application A, and further detection of user B's historical behavioral characteristics data regarding the ads can be performed. For instance, detection of user B's historical behavioral characteristics data on other applications. Here, historical behavior refers to the historical behavior describing object B's interest in the ads.

[0065] In one embodiment, considering that in practical applications there may be objects that have a need for content promotion, but because this need has not been captured or identified by the content promotion system, the system has not previously promoted content to these objects. This results in the inability to accurately calculate the object's behavioral characteristics in response to the promoted content, thus preventing the user from becoming a target audience for content promotion. Therefore, to address this situation, the content push evaluation model of the content promotion system can be adjusted to avoid missing potential target audiences. Specifically, the step "obtaining the object's behavioral information in response to the promoted content" may include: Adjust the content promotion evaluation threshold in the content promotion system; Based on the adjusted content promotion evaluation threshold, promotional content is sent to the target audience; Obtain information about the target's behavior towards the promoted content within a preset time interval.

[0066] The content promotion evaluation threshold is the threshold for determining when to push content to a target audience. For example, the content promotion evaluation threshold can be the threshold used by the content promotion system to determine when to push content to a target audience. Specifically, content promotion evaluation can be used to assess the value generated by pushing promotional content to a target audience. For example, if the promotional content is an advertisement and the target audience is a user, then content promotion evaluation is advertising promotion evaluation. Specifically, it can be used to assess the value generated by pushing an advertisement to a user, or exposing an advertisement to a user.

[0067] In practical applications, to avoid disturbing users and improve content promotion efficiency, it can be set that if the content promotion evaluation result is lower than the content promotion evaluation threshold, the promotion will not be executed. Taking advertising as an example, if the content promotion evaluation result for promoting ad C to user B is lower than the content promotion evaluation threshold, the promotion will not be executed, meaning ad C will not be exposed to user B.

[0068] Similarly, taking advertising as an example, in actual applications, there are some users whose estimated advertising promotion evaluation results are very low. As a result, although there are users' advertising requests, no ads are returned, resulting in no exposure and clicks. Therefore, it is impossible to determine whether these users are silent users. In this case, users who meet the criteria for silent users will become hidden silent users.

[0069] There are several ways to adjust the content promotion evaluation threshold. For example, it can be adjusted by lowering the original content promotion evaluation threshold; or by setting the original content promotion evaluation threshold to a content promotion evaluation range; and so on.

[0070] In one embodiment, see Figure 3 One could increase the ad exposure of latent users by relaxing the ad promotion evaluation threshold, that is, by lowering the ad promotion evaluation threshold, thereby prompting latent users to become silent users.

[0071] After adjusting the content promotion evaluation threshold of the content promotion system, promotional content can be sent to the target based on the adjusted content promotion evaluation threshold, thereby obtaining the target's behavior information regarding the promotional content within a preset time interval.

[0072] 104. When the object has the aforementioned historical behavior feature data, content recall is performed on the set of promotional content based on the historical behavior feature data to obtain the object's target set of promotional content.

[0073] Content recall refers to the process of selecting promotional content from a collection of promotional content. Specifically, in this application, content recall can be used to select corresponding promotional content from the collection of promotional content and add it to the target promotional content collection of the object. For example, in the example of advertising, content recall refers to the process of selecting corresponding advertisements from the collection of advertisements and adding them to the user's target advertisement collection.

[0074] The target promotional content set is a collection of promotional content recalled from the promotional content set. For example, in the advertising example, the target promotional content set is a collection of advertisements recalled from the advertising set. It is worth noting that in this application, there can be multiple methods for recalling content from the promotional content set based on historical behavioral characteristic data. Therefore, the target promotional content set of the object can be generated by adding promotional content recalled through different methods to the object's target promotional content set.

[0075] There are several ways to recall promotional content based on historical behavioral data. For example, in the case of advertisements, ads that inactive users have clicked on on other sites can be selected from the ad set and added to the target ad set; ads that inactive users clicked before becoming inactive can be added to the target ad set; ads that inactive users' social network users have clicked can be added to the target ad set, where social network users can be users who are related to the inactive user, such as the inactive user's social friends or contacts in the inactive user's address book; and so on.

[0076] There are various ways to recall content based on historical behavioral feature data for a set of promoted content. In another embodiment, not only historical behavioral feature data can be considered, but also the content promotion evaluation model for the promoted content can be taken into account. Content recall can be achieved by adjusting the original content promotion evaluation model. For example, the content promotion evaluation model can be adjusted by adding input features to the original model, so that the adjusted model can determine the content promotion evaluation result corresponding to the target by referring to the added features. Specifically, the step "recalling content based on historical behavioral feature data for a set of promoted content to obtain the target set of promoted content for the target" can include: Determine the original promotion evaluation model for the promoted content; Feature extraction is performed on historical behavioral data to obtain the content interaction characteristics of the target in response to the promoted content; Add content interaction features to the original promotion evaluation model so that the promotion evaluation model can determine the content promotion evaluation result corresponding to the target by referring to the content interaction features. Based on the content promotion evaluation results, content recall is performed on the set of promotional content to obtain the target set of promotional content for the target audience.

[0077] The original promotion evaluation model refers to the original content promotion evaluation model, which is used to evaluate the promotion effect when promoting content to the target audience. For example, the value generated when promoting content to the target audience can be used as the promotion effect.

[0078] There are various ways to implement a content promotion evaluation model. For example, a content promotion evaluation model can be a linear model, a non-linear model, or a combination of linear and non-linear models. For instance, a non-linear model can include a neural network model.

[0079] In one embodiment, taking advertising as an example, the expected cost per mile (eCPM) can be used as the content promotion evaluation result. Therefore, the original content promotion evaluation model can be: eCPM = pCTR × pCVR × bid × adjustment factor + quality eCPM.

[0080] CTR (Click-Through Rate) is the click-through rate, specifically defined as clicks / impressions. pCTR (Predicted Click-Through Rate) is the predicted click-through rate, calculated by the click-through rate prediction model in the advertising system for each ad in the ad set. pCVR (Predicted Conversion Rate) is the predicted conversion rate, calculated by the conversion rate prediction model in the advertising system for each ad in the ad set. eCPM is the expected revenue per thousand impressions, calculated based on ad bids, pCTR, and pCVR, serving as the ranking criterion for the advertising system and directly linked to advertising revenue. qualityeCPM is a weighted eCPM considered from the perspective of long-term ecosystem impact, providing weights for high-quality ads. Common quality eCPMs can include those weighted based on pCTR, those weighted based on pCVR, etc.

[0081] Among them, content interaction features describe the characteristics of an object's content interaction behavior in response to promotional content. For example, content interaction behavior may include content browsing behavior, content liking behavior, content commenting behavior, content product purchase behavior, etc.

[0082] There are various ways to extract features. For example, one can generate corresponding vectors based on historical behavioral feature data through vectorization and use these vectors as the content interaction features of the object. Another example is to train a neural network model for feature extraction and use historical behavioral feature data as the model input to generate the content interaction features of the object through the neural network model.

[0083] In the example of the advertisement, the object can be a dormant user. The historical behavioral characteristics data of the object can include the dormant user's historical ad click behavior, historical ad comment behavior, historical ad collection behavior, historical ad product purchase behavior, etc.

[0084] In this example, such as Figure 3 As shown in Figure 1001, feature extraction can be performed on the historical behavioral data of inactive users in response to advertisements to extract content interaction features of inactive users. Among them, the extracted content interaction features can include features of the information flow context, such as the features of articles and videos viewed by inactive users on the information flow side; the extracted content interaction features can include features of the advertisement context, such as the information features of inactive users' historical advertisement exposure and swiping; in addition, the extracted content interaction features can also include the behavioral features of inactive users before and after becoming inactive, and so on.

[0085] Furthermore, the extracted content interaction features can be added to the original promotion evaluation model to adjust it. For example, the extracted content interaction features can be added as input data to the original promotion evaluation model, allowing the model to understand the relationship between these objects and the promotional content, thereby improving the model's recognition accuracy and correcting errors. Optionally, in practical applications, the content interaction features can be updated periodically to update the input of the content promotion evaluation model.

[0086] In the example of advertising, the content promotion evaluation model may include a neural network model for calculating pCTR. Based on the original input data of this neural network model, the extracted content interaction features can be used as new input data for the neural network model to add the extracted content interaction features to the original content promotion evaluation model, thereby adjusting the original content promotion evaluation model.

[0087] In this application, after adjusting the original promotion evaluation model of the promotion content to obtain the adjusted promotion evaluation model, the content promotion evaluation result of each promotion content in the promotion content set corresponding to the object can be calculated based on the adjusted promotion evaluation model.

[0088] The content promotion evaluation result is used to measure the promotion effect when promoting content to an audience. Specifically, after calculating the promotion evaluation result for the audience in response to the promotion content, the promotion effect when the promotion content is promoted to that audience can be estimated. In the advertising example, eCPM can be used as the content promotion evaluation result. For example, by calculating the eCPM of dormant user D for ad E, the value generated by exposing ad E to user D can be estimated.

[0089] Furthermore, based on the calculated promotion evaluation results, corresponding promotional content can be selected from the promotional content set and added to the target promotional content set of the object to achieve content recall of the promotional content set and obtain the target promotional content set of the object.

[0090] Based on the promotion evaluation results, there are several ways to select relevant promotional content from the promotional content set. For example, the promotional content can be sorted based on the evaluation results to obtain sorted promotional content. Furthermore, a preset number or preset proportion of promotional content can be selected from the sorted content; another example is selecting promotional content within a preset sequence range from the sorted content; and so on. Furthermore, the selected promotional content can be added to the target promotional content set of the object.

[0091] In another embodiment, when the object possesses the historical behavioral feature data, the object's preference information for promotional content can be determined based on the historical behavioral feature data. This allows for content retrieval based on the preference information for the set of promotional content. Specifically, the step "retrieving content based on the historical behavioral feature data for the set of promotional content to obtain the object's target set of promotional content" may include: Based on historical behavioral data, determine the target audience's preference information for the promoted content; Based on preference information, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the target audience.

[0092] Among them, the object's preference information for the promoted content is relevant information used to describe the object's preference for the promoted content. For example, the preference information can be determined based on the object's historical behavior towards the content, such as the target object's historical search behavior, historical click behavior, historical viewing behavior, historical purchase behavior, historical collection behavior, etc.

[0093] Preference information can take many forms. For example, preference information can exist in the form of interest tags; or in the form of interest quality scores; and so on.

[0094] Content recall refers to the process of selecting promotional content from a set of promotional content. There are various ways to recall promotional content based on preference information. For an example of advertising, please refer to... Figure 3 As shown in Figure 1002, interest ads are recalled from the ad set based on the preceding media information, related ads are recalled from the ad set based on the preceding cross-site ad information, and related ads are recalled from the ad set based on interest quality scores.

[0095] In one embodiment, promotional content matching an object's interest tags can be selected from a set of promotional content, and the selected content can be added to the object's target promotional content set. For example, based on an object's interest tags, the target industries and products the object is interested in can be determined, and advertisements matching the target industry and the target product can be selected from an advertisement set and added to the user's target advertisement set. Specifically, when applied to advertising, this process is as follows: Figure 3 The process shown in Figure 1002 involves recalling interest-based ads from a historical ad set based on media information.

[0096] In another embodiment, the object group to which the object belongs can be determined based on the object's preference information for the promoted content. This enables content retrieval of the promoted content set by collaborating with the object group to which the object belongs, thereby obtaining the object's target promoted content set. Specifically, the step "retrieving content from the promoted content set based on preference information to obtain the object's target promoted content set" may include: Based on preference information, determine the object group to which the object belongs; Select promotional content that is associated with the target group from the collection of promotional content; The selected promotional content is sorted to obtain the sorted promotional content; Based on the ranking results, content recall is performed on the ranked promotional content to obtain the target promotional content set for the object.

[0097] Here, an object group is a collection of objects consisting of at least one object. For example, if the object is a user, then the object group can be a collection of users that includes at least one user.

[0098] There are several ways to determine the object group to which an object belongs based on preference information. For example, the object group to which an object belongs can be determined based on different forms of preference information.

[0099] In one embodiment, the preference information can be in the form of interest tags. Based on the object's interest tags for the promoted content, the object group to which the object belongs can be determined. For example, objects with common selection tags can be selected to generate an object group. Based on the object's interest tags, objects with common interest tags with the object can be determined, and thus the object group to which the object belongs can be determined.

[0100] In another embodiment, the preference information can take the form of an interest quality score. Based on the object's interest quality score for different types of promotional content, the object group to which the object belongs can be determined. For example, objects with similar interest quality scores for the same type of promotional content can be selected to generate an object group. Then, based on the object's interest quality score for different types of promotional content, objects with similar interest quality scores for the same type of promotional content can be identified, and thus the object group to which the object belongs can be determined. It is worth noting that the similarity metric can be set based on business needs.

[0101] Furthermore, promotional content associated with the target group can be selected from the collection of promotional content. There are various ways to associate these contents. For example, promotional content viewed by any member of the target group can be used as promotional content associated with that target group; alternatively, promotional content viewed by members of the target group can be sorted based on statistical information about the promotional content, and a portion of the promotional content can be selected from the sorted results as promotional content associated with that target group; and so on.

[0102] Furthermore, the selected promotional content can be sorted. There are various sorting methods available, such as sorting based on statistical information about the promotional content. Specifically, in the advertising example, the selected ads can be sorted based on their CTR (Click-Through Rate); or based on their CVR (Cost-Return Rate); and so on.

[0103] Based on the sorting results, there are several ways to determine the corresponding promotional content from the sorted promotional content. For example, a preset number or preset proportion of promotional content can be selected from the sorted content; another example is selecting promotional content within a preset sequence number range from the sorted content; and so on. Furthermore, the selected promotional content can be added to the target promotional content set of the object.

[0104] In another embodiment, content retrieval based on preference information for a set of promotional content can be achieved by combining the object's preference information for promotional content with the content description information of the promotional content. Specifically, the step "retrieving content based on preference information for a set of promotional content to obtain the object's target set of promotional content" may include: Obtain the content description information of the promotional content; Based on preference information and content description information, content preference matching is performed between the target and the promotional content; Based on the matching results, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the object.

[0105] The content description information of the promotional content is information that describes the promotional content. For example, the promotional content can be described in the form of tags; or it can be described in the form of images; and so on.

[0106] Content preference matching is used to determine the degree of preference of an object for the promoted content. Specifically, by performing content preference matching between an object and the promoted content, it can be determined how much interest the object has in the promoted content.

[0107] There are multiple ways to match target audiences with promotional content based on preference information and content description information. Specifically, the method can be determined based on the form of preference information and content description information.

[0108] In one embodiment, both preference information and content description information can be in the form of tags. Therefore, an object can include at least one tag representing preference information, and similarly, promotional content can include at least one tag representing content description information. Thus, content preference matching between the object and promotional content can be performed based on the tags they share. For example, the content preference matching result between the object and the promotional content can be determined based on the number of tags they share. Alternatively, vectorization techniques can be used to convert the tags of the object and the promotional content into corresponding vectors, and the content preference matching result between the object and the promotional content can be determined by calculating the distance between the vectors; and so on.

[0109] In another embodiment, both preference information and content description information can be in the form of images. The content preference matching result between the object and the promoted content can be determined based on the calculation of the images. For example, the content preference matching result can be determined by calculating the image relevance; or by calculating the image similarity; and so on.

[0110] Furthermore, based on the matching results, content retrieval can be performed on the set of promotional content. For example, based on the matching results, relevant promotional content can be selected from the set of promotional content and added to the target set of promotional content. For example, based on the matching results, the promotional content in the set of promotional content can be sorted, and relevant promotional content can be selected from the sorted content and added to the target set of promotional content. For instance, a preset number or preset proportion of promotional content can be selected from the sorted content; or, promotional content within a preset sequence range can be selected from the sorted content; and so on.

[0111] 105. When the object does not have historical behavioral feature data, sort the promotion content set to recall the object's promotion content and obtain the object's target promotion content set.

[0112] The process of sorting the promotional content set refers to ranking the promotional content within the set to obtain the sorted content. It is worth noting that there are multiple ways to sort the promotional content set in this application. Therefore, the target promotional content set of the object can be generated by adding promotional content retrieved through different methods to the object's target promotional content set.

[0113] There are several ways to sort the promotional content set to recall the target's promotional content. For example, it can be achieved through sorting exploration. Specifically, the step "sorting the promotional content set to recall the target's promotional content, obtaining the target promotional content set for the target" can include: Obtain promotion statistics for the promoted content, as well as the target audience's preference information for the promoted content; Based on promotion statistics and preference information, determine the content promotion evaluation parameters for the promoted content; Based on content promotion evaluation parameters, the promotion content set is sorted to recall the target promotion content and obtain the target promotion content set for the target.

[0114] The promotion statistics for the promoted content refer to relevant information collected during the promotion of the content. For example, the promotion statistics for the promoted content may include statistics on the interactive behavior of the target audience in response to the promoted content. The statistical information itself is information obtained through statistical analysis; for example, the statistical information may include totals, averages, medians, modes, maximum values, minimum values, etc.

[0115] Taking advertising as an example, advertising promotion statistics can include statistical information obtained after statistical analysis of advertising interaction behaviors such as ad clicks, ad comments, ad favorites, and ad product purchases. For example, promotion statistics can include CTR, CVR, and other advertising promotion statistics.

[0116] Among them, the object's preference information for the promoted content is relevant information used to describe the object's preference for the promoted content. For example, the preference information can be determined based on the object's historical behavior towards the content, such as the object's historical search behavior, historical click behavior, historical viewing behavior, historical purchase behavior, historical collection behavior, etc.

[0117] Preference information can take many forms. For example, it can exist as interest tags, which can be used to characterize an object's interest in content related to those tags, thus determining that the object has a preference for promotional content related to those tags. Alternatively, preference information can exist as interest quality scores. Specifically, models can be built based on an object's preference information to predict the object's interest quality score for different types of promotional content. For example, it can be used to predict the object's interest quality score for promotional content in different industries, or for promotional content for different products; and so on.

[0118] In the example of the advertisement, the interest tags of silent users for the advertisement can be used to characterize whether silent users are interested in content related to that interest tag. For example, if the interest tag of a silent user for the advertisement is "pet", it means that the silent user is interested in pet-related content, and thus it can be determined that the silent user has a preference for pet-related advertisements.

[0119] Among them, the content promotion evaluation parameters are parameters used to assist in the evaluation of content promotion. Specifically, based on the content promotion evaluation parameters, the promotion content in the promotion content set can be sorted so that the content promotion evaluation result can be determined according to the sorting result.

[0120] There are several ways to determine content promotion evaluation parameters based on promotion statistics and preference information. For example, the primary evaluation parameter for the promoted content can be determined based on promotion statistics. As an example, if the promotion statistics are CRT (Cost-to-Rate) data, then pCTR (Promotional Criterion Rate) can be used as the primary evaluation parameter for the promoted content; similarly, if the promotion statistics are CVR (Cost-to-Volume Ratio) data, then pCVR (Promotional Criterion Rate) can be used as the primary evaluation parameter for the promoted content; and so on.

[0121] For example, a second evaluation parameter for the promoted content can be determined based on preference information. As an example, preference information could be the target audience's interest quality score regarding the promoted content, which could then be used as the second evaluation parameter for the promoted content.

[0122] Furthermore, content promotion evaluation parameters can be determined based on the first evaluation parameter and the second evaluation parameter. For example, content promotion evaluation parameters can be selected from the first evaluation parameter and the second evaluation parameter; or both the first evaluation parameter and the second evaluation parameter can be used as content promotion evaluation parameters; and so on.

[0123] After determining the content promotion evaluation parameters, the promotion content set can be sorted based on these parameters to recall the target promotion content and obtain the target promotion content set for the target. Specifically, the step "sorting the promotion content set based on the content promotion evaluation parameters to recall the target promotion content and obtain the target promotion content set for the target" can include: Determine the original promotion evaluation model for the promoted content; Add content promotion evaluation parameters to the original promotion evaluation model so that the promotion evaluation model after adding the parameters can determine the content promotion evaluation result of the target audience based on the content promotion evaluation parameters. Based on the content promotion evaluation results, the promotion content set is sorted. Based on the ranking results, the target promotion content of the object is retrieved.

[0124] Once the content promotion evaluation parameters for the promoted content are determined, these parameters can be added to the content promotion evaluation model to adjust the model. In this way, the process of evaluating content promotion through the adjusted model is the process of using the content promotion evaluation parameters to assist in the evaluation.

[0125] In one embodiment, the content promotion evaluation model can be: eCPM = pCTR × pCVR × bid × adjustment factor + Furthermore, the step "Add content promotion evaluation parameters to the content promotion evaluation model to adjust the content promotion evaluation model" can provide... Figure 3 The ranking exploration process is shown in Figure 1003. As an example, if the content promotion evaluation parameter can be pCTR, then adding the content promotion evaluation parameter to the content promotion evaluation model to adjust the model can be specifically as follows: eCPM' = eCPM + alpha × pCTR, where alpha is the coefficient term corresponding to pCTR. In this way, by weighting only requests from inactive users, support efficiency can be maximized while keeping costs under control. As another example, if the content promotion evaluation parameter can be the interest quality score S, then adding the content promotion evaluation parameter to the content promotion evaluation model to adjust the model can be specifically as follows: eCPM' = eCPM + beta × S, where beta is the coefficient term corresponding to S.

[0126] In this application, the content promotion evaluation parameters are added to the original promotion evaluation model to obtain the post-added promotion evaluation model. Based on the post-added promotion evaluation model, the content promotion evaluation results of each promotion content in the promotion content set corresponding to the object can be calculated.

[0127] Among them, the content promotion evaluation result is used to measure the promotion effect when the promotion content is promoted to the target audience. Specifically, after calculating the promotion evaluation result of the target audience for the promotion content, the promotion effect when the promotion content is promoted to the target audience can be estimated.

[0128] In the example of advertising, eCPM can be used as a promotion evaluation result. For example, by calculating the eCPM of dormant user D for ad E, the value that can be generated when ad E is exposed to user D can be estimated.

[0129] Furthermore, based on the calculated content promotion evaluation results, the promotional content in the promotional content set can be sorted, and based on the sorting results, corresponding promotional content can be selected from the promotional content set and added to the target promotional content set of the object. For example, a preset number or preset proportion of promotional content can be selected from the sorted promotional content; or, promotional content within a preset sequence range can be selected from the sorted promotional content, etc. Furthermore, the selected promotional content can be added to the target promotional content set of the object.

[0130] The process of sorting the promotional content set to retrieve the target promotional content can take many forms. For example, it can be achieved through popular content retrieval. In the advertising example, it can retrieve ads for specific user segments, popular ads from various industries, local ads for inactive users, etc. Specifically, the step "sorting the promotional content set to retrieve the target promotional content set" can include: Obtain the content attribute information of the promotional content, as well as the object attribute information of the object; Based on content attribute information and object attribute information, the promotion content set is sorted to recall the object's promotion content and obtain the object's target promotion content set.

[0131] Among them, the content attribute information of the promotional content refers to relevant information describing the attributes of the promotional content. For example, the content attribute information may include the geographical affiliation information, industry affiliation information, product information involved, target audience information, etc.

[0132] Among them, the object attribute information of an object is related information describing the attributes of the object. For example, taking a user as an example, the object attribute information of an object may include the user's gender information, age information, geographical location information, occupation information, etc.

[0133] There are several ways to sort a collection of promotional content based on content attribute information and object attribute information. For example, the content can be sorted based on its geographic location information and the object's geographic location information. For instance, the higher the correlation between the geographic location information of the promotional content and the geographic location information of the object, the higher the weight of that promotional content in the content sorting process. Specifically, using advertising as an example, see [link to relevant documentation]. Figure 3In section 1004, since local ads are generally placed by local "small but beautiful" businesses, they can bring a certain sense of novelty to users, stimulate their interest, and efficiently promote clicks and conversions. Therefore, the ads in the ad set can be sorted based on the ad's geographic affiliation information and the geographic location information of the dormant users, so that the local ads of dormant users can have a higher weight. In this way, the local ads of dormant users can be selected from the ad set and added to the target ad set of dormant users.

[0134] For example, the promotional content can be sorted based on its industry affiliation. Specifically, promotional content belonging to different industries can be sorted to obtain the ranking results for each industry, and the top-ranked promotional content from each industry can be retrieved and added to the target promotional content set. See the example of advertising. Figure 3 In option 1004, based on the industry affiliation information of the promotional ads, popular promotional content from each industry can be selected from the ad collection and added to the target ad collection of the object.

[0135] For example, based on an object's attribute information, the target audience corresponding to the object can be determined, and popular promotional content from the expected target audience can be selected from the promotional content set and added to the object's target promotional content set. Specifically, using advertising as an example, see [link to relevant documentation]. Figure 3 In the case of 1004, based on the age, gender, and occupation information of inactive users, popular ads for the corresponding segment of the user can be selected from the ad collection and added to the user's target ad collection.

[0136] It is worth noting that there can be multiple ways to measure popularity. For example, popular promotional content can be determined by sorting the promotional content and then using the sorting results. Specifically, the step "based on content attribute information and object attribute information, sorting the set of promotional content to recall the object's promotional content and obtain the object's target promotional content set" can include: Based on content attribute information and object attribute information, select promotional content to be sorted from the promotional content set; Sort the promotional content to be sorted to obtain the sorted promotional content; Based on the ranking results, the target promotion content of the object is retrieved from the ranked promotion content to obtain the target promotion content set of the object.

[0137] There are various sorting methods. For example, sorting can be based on the publication time of the promoted content; or it can be based on the historical promotion performance of the promoted content; and so on. Taking advertising as an example, the historical promotion performance of an advertising can be determined based on its exposure rate, click-through rate, like rate, collection rate, comment rate, etc., and then the advertising can be sorted based on its historical promotion performance.

[0138] Similarly, there are several ways to select promotional content from the sorted promotional content. For example, a preset number or proportion of promotional content can be selected from the sorted content; or promotional content within a preset sequence number range can be selected from the sorted content; and so on. Furthermore, the selected promotional content can be added to the target promotional content set of the object.

[0139] 106. Determine the target promotion content for the target from the target promotion content set, and send the target promotion content to the target.

[0140] There are several ways to determine the target promotion content of an object from the target promotion content set. For example, a content promotion evaluation model can be used to calculate the content promotion evaluation results of each promotion content. Based on the calculation results, the promotion content in the target promotion content set can be sorted to obtain the sorted promotion content, and the target promotion content of the object can be determined based on the sorting results. Specifically, the step "determine the target promotion content of the object from the target promotion content set" may include: The content promotion evaluation results of the promoted content are calculated using a content promotion evaluation model. Based on the calculation results, the target promotion content for the object is determined from the set of target promotion content.

[0141] The content promotion evaluation model can include the original promotion evaluation model or the adjusted content promotion evaluation model. The adjusted content promotion evaluation model can include the post-added promotion evaluation model obtained by adding content promotion evaluation parameters to the original promotion evaluation model, or the post-added promotion evaluation model obtained by adding content interaction features to the original promotion evaluation model, and so on.

[0142] For example, a content promotion evaluation model can be used to calculate the evaluation results of the promoted content, and based on the calculation results, the promoted content in the set can be sorted to obtain the sorted promoted content. Furthermore, target promoted content for an object can be determined from the sorted promoted content; for example, a preset number or preset proportion of promoted content can be selected from the sorted promoted content; or, for example, promoted content within a preset sequence range can be selected from the sorted promoted content; and so on.

[0143] For example, when determining the target promotional content for an object from the sorted promotional content, it can specifically include two stages: coarse ranking and fine ranking. Furthermore, during the coarse ranking process, a certain percentage of slots can be reserved for candidate promotional content that has a high content preference match with the target object, thereby increasing the probability of selecting that candidate promotional content during the fine ranking. For instance, the step "Based on the calculation results, determine the target promotional content for the target object from the set of candidate promotional content" can include: Obtain the content preference matching results between candidate promotional content and target audience; determine the weight information corresponding to the candidate promotional content based on the content preference matching results; and determine the target promotional content for the target audience from the candidate promotional content set based on the weight information.

[0144] In this application, after determining the target promotion content for an object from the target promotion content set, the target promotion content can be sent to the object. For example, the server can determine the target promotion content for an object from the target promotion content set through a content promotion system, and then send the target promotion content to the object. In the example of advertising, the server can determine the target advertisement for a dormant user from the target advertisement set through an advertisement promotion system, and send the target advertisement to the dormant user to recommend advertisements of interest to the dormant user.

[0145] Furthermore, considering that in practical applications, if an object's behavioral characteristics towards the promoted content meet the preset negative behavioral conditions, there could be two reasons: First, the object has a need for content promotion, but because this need has not been captured or identified by the content promotion system, the system has not promoted content to the object in the past, thus making it impossible to accurately calculate the object's behavioral characteristics towards the promoted content, thereby preventing the user from becoming a target of content promotion; second, the object is naturally not interested in the promoted content. Therefore, in this application, after determining the target promoted content for the object and sending the target promoted content to the object, the object type of the object towards the promoted content can be further evaluated through the object's target behavioral characteristics towards the target promoted content. Specifically, the method for processing the promoted content may also include: Determine the target behavioral characteristics of the target object in relation to the target promotional content; Based on the target behavioral characteristics and the preset negative behavioral conditions, determine the target audience type for the promotional content.

[0146] There are several ways to determine the target behavioral characteristics of an object in relation to the target promotional content. For example, one can obtain the target behavioral information of the object in relation to the target promotional content; and then perform statistical analysis on the target behavioral information to obtain the target behavioral characteristics of the object in relation to the target promotional content. For specific implementation methods, please refer to the steps "obtain the behavioral information of the object in relation to the promotional content; perform statistical analysis on the behavioral information to obtain the behavioral characteristics of the object in relation to the promotional content", which will not be elaborated here.

[0147] The "object type" in this context describes whether an object is a user who is inherently uninterested in the promoted content. Specifically, corresponding to the reasons for meeting the preset negative behavior conditions, the object type can include the following two types: First, an object type that is interested in the promoted content. For this type of object, the user is interested in the promoted content, but because the content promotion system has not captured the user's true interest, the system cannot recommend promoted content that the user is interested in, resulting in a negative reaction to the recommended content. The second type is an object type that is indifferent to the promoted content. For this type of object, the user naturally forms a barrier to the promoted content, resulting in a negative reaction to the recommended content. Negative reactions to promoted content can include not clicking on the exposed promoted content, turning off the exposed promoted content within a predetermined time threshold, and blocking the promoted content, etc. Therefore, after determining the target behavioral characteristics of the target audience in relation to the targeted promotional content, the target audience type can be determined based on these characteristics and preset negative behavioral conditions. For example, if the target behavioral characteristics meet the preset negative behavioral conditions, the target audience is determined to be indifferent to the promotional content; if the target behavioral characteristics do not meet the preset negative behavioral conditions, the target audience is determined to be responsive to the promotional content.

[0148] See the example of the advertisement. Figure 3 The method for processing promotional content described in this application can be used to determine the target advertisement for a silent user from the target advertisement set and send the target advertisement to the silent user to explore the conversion of silent users into ordinary users. Ordinary users refer to users who are interested in the promotional content, that is, users who are interested in the advertisement.

[0149] See in this example. Figure 3 It can be seen that some silent users become ordinary users and exit exploration after activation, while some silent users remain silent after exploration. This indicates that these silent users are not interested in any advertisements and will not bring revenue to the platform. Therefore, in the short term, the exposure of advertisements can be reduced to reduce the disturbance to users.

[0150] As can be seen from the above, this embodiment can obtain the object's promotional content request; determine the promotional content set based on the promotional content request, wherein the promotional content set includes at least one promotional content; if the object's behavioral characteristics toward the promotional content meet the preset negative behavioral conditions, then detect the object's historical behavioral characteristic data toward the promotional content; when the object has historical behavioral characteristic data, perform content recall on the promotional content set based on the historical behavioral characteristic data to obtain the object's target promotional content set; when the object does not have historical behavioral characteristic data, perform content sorting processing on the promotional content set to recall the object's promotional content to obtain the object's target promotional content set; determine the object's target promotional content from the target promotional content set, and send the target promotional content to the object.

[0151] This solution targets individuals whose behavioral characteristics in response to promoted content meet preset negative behavioral conditions. After receiving a request for promoted content from a target individual, it determines how to retrieve relevant content from the promoted content set based on whether the target individual has historical behavioral data related to the promoted content. This allows for precise content promotion tailored to the target individual. Specifically, if the target individual has historical behavioral data related to the promoted content, the solution can retrieve content based on this historical data. This allows for accurate prediction of the promoted content that the target individual might be interested in by referencing their content interaction history with the promoted content.

[0152] Furthermore, if the target audience lacks historical behavioral data regarding the promoted content, this solution can retrieve the corresponding promoted content from the promoted content set through sorting exploration, thereby increasing the probability of predicting the promoted content that the target audience is interested in. Therefore, when promoting content to a target audience, this solution can improve the accuracy and efficiency of promotion by increasing the processing efficiency of promoted content.

[0153] Furthermore, when applied to advertising scenarios, this solution activates dormant users through various methods, increasing interaction between users who haven't clicked and the advertising system, thereby improving the accuracy of the advertising system's recommendations. Therefore, this solution can increase the probability of exposing users to ads they are interested in, enhancing user experience and, in the long run, benefiting user activity and retention. For media outlets, this solution can improve user click-through conversion efficiency, increasing media traffic GMV. For advertising systems, this solution can increase positive click and conversion examples generated by users within the system, helping to improve the system's ability to identify users' genuine commercial interests.

[0154] Based on the method described in the above embodiments, the following examples will provide further detailed explanations.

[0155] In this embodiment, the integration of the promotional content processing device into the server and terminal will be used as an example. The server can be a single server or a server cluster composed of multiple servers; the terminal can be a mobile phone, tablet computer, laptop computer or other devices.

[0156] like Figure 4 As shown, a method for processing promotional content is described below, with the specific process as follows: 201. The terminal sends a request for promotional content for an object to the server.

[0157] In one embodiment, the method for processing promotional content described in this application can be applied to a search application, wherein the promotional content can be an advertisement, and the target audience can be users of the search application. Since a search application can not only provide search results to users but also display related advertisements while showing search results to promote content to users, when a user generates a search request in the search application, a corresponding advertisement request can be generated, and the terminal can send both the search request and the advertisement request to the server.

[0158] 202. The server retrieves the promotional content request for the object.

[0159] 203. The server determines the set of promotional content based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content.

[0160] In one embodiment, the set of promotional content determined by the server based on the promotional content request may include a set of historical promotional content and a set of content to be promoted. The set of historical promotional content may include at least one historical promotional item, and the set of content to be promoted may include at least one item to be promoted. For example, in the case of an advertisement, the set of advertisements determined by the server based on the advertisement request may include a set of historical advertisements and a set of advertisements to be promoted.

[0161] 204. If the object's behavioral characteristics toward the promotional content meet the preset negative behavioral conditions, the server will detect the object's historical behavioral characteristic data toward the promotional content.

[0162] In the example of the advertisement, the preset negative behavior condition can be specifically defined as: the advertisement has been viewed more than 60 times in the same application within the last 30 days, and there has been no advertisement click behavior. If a user's behavior characteristics towards the advertisement meet the preset negative behavior condition, the user can be identified as a dormant user of the advertisement in that search application. Furthermore, the historical behavioral characteristic data of dormant users towards the advertisement can be detected. For example, historical behavioral characteristic data can refer to the context of the dormant user in media information and cross-media context.

[0163] It is worth noting that this application does not limit the setting method of preset negative behavioral conditions, which can be expanded or simplified according to the actual application scenario. Furthermore, this invention does not limit the form of historical behavioral feature data, which can be expanded or simplified according to the actual application scenario.

[0164] 205. When an object has historical behavioral characteristic data, the server performs content recall based on the historical behavioral characteristic data for the set of promotional content to obtain the target set of promotional content for the object.

[0165] There are various ways to recall content based on historical behavioral data for a set of promoted content. For example, in the case of advertising, the server can determine the interests of dormant users based on historical behavioral data, enabling interest-based recall of the entire ad set. Another example is that the server can extract a content interaction model based on historical behavioral data and adjust the content promotion evaluation model by adding content interaction features. This adjusted model can then be used to recall ads from the ad set. For instance, interest-based recall can include recalling ads that dormant users are interested in based on the context of media information, recalling relevant ads based on cross-site ad context, and recalling relevant ads based on interest segments. Content interaction features can include features from the context of news feeds and ads.

[0166] 206. When the object does not have historical behavioral feature data, the server performs content sorting processing on the set of promotional content to recall the object's promotional content and obtain the object's target set of promotional content.

[0167] There are various ways to sort the promotional content set. For example, in the case of advertisements, the server can sort the advertisements in the ad set based on their popularity and recall the most popular advertisements. Alternatively, the server can add content promotion evaluation parameters to the content promotion evaluation model to adjust the model, obtaining an adjusted model, and then use the adjusted model to sort the advertisements in the ad set to recall advertisements based on the sorting results. For example, popular advertisement recall can include popular advertisements segmented by audience, popular advertisements segmented by industry, and local advertisements. It is worth noting that this application does not limit the specific statistical criteria for popular advertisements and can expand or simplify them according to the actual application scenario.

[0168] 207. The server determines the target promotion content for an object from the target promotion content set and sends the target promotion content to the object's terminal.

[0169] 208. The terminal receives the target promotional content and displays the target promotional content to the target audience.

[0170] As can be seen from the above, this application embodiment can target objects whose behavioral characteristics for the promoted content meet preset negative behavioral conditions. After obtaining the promoted content request from the target object, it determines how to recall the promoted content for the target object from the promoted content set based on whether the target object has historical behavioral characteristic data for the promoted content, so as to achieve precise content promotion for the target object. Specifically, if the target object has historical behavioral characteristic data for the promoted content, this application embodiment can recall content based on the historical behavioral characteristic data. In this way, by referring to the target object's content interaction history for the promoted content, it is possible to accurately predict the promoted content that the target object is interested in. If the target object does not have historical behavioral characteristic data for the promoted content, this application embodiment can recall the corresponding promoted content from the promoted content set by sorting and exploring, so as to increase the probability of predicting the promoted content that the target object is interested in. Therefore, when promoting content to the target object, this application embodiment can improve the accuracy and efficiency of promotion by improving the processing efficiency of the promoted content.

[0171] Furthermore, when applied in advertising scenarios, this solution is used to activate dormant users. Specifically, it can be achieved through strategies such as user interest mining, popular ad recall, and ranking formula optimization, in order to increase the click-through rate of dormant users on ads and the system's GMV.

[0172] To better implement the above methods, this application also provides a device for processing promotional content, which can be integrated into a server or a terminal. The server can be a single server or a server cluster consisting of multiple servers; the terminal can be a mobile phone, tablet computer, laptop computer, or other devices.

[0173] For example, such as Figure 5 As shown, the processing device for the promotional content may include an acquisition unit 301, a set determination unit 302, a detection unit 303, a first recall unit 304, a second recall unit 305, and a content determination unit 306, as follows: The acquisition unit 301 can be used to acquire the promotional content request of an object; The set determination unit 302 can be used to determine a set of promotional content based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content; The detection unit 303 can be used to detect the historical behavioral feature data of the object in relation to the promotional content if the behavioral features of the object in relation to the promotional content meet the preset negative behavioral conditions. The first recall unit 304 can be used to recall content based on the historical behavior feature data of the object when the object has the historical behavior feature data, so as to obtain the target promotion content set of the object. The second recall unit 305 can be used to sort the promotional content set when the object does not have the historical behavior feature data, so as to recall the promotional content of the object and obtain the target promotional content set of the object. The content determination unit 306 can be used to determine the target promotion content of the object from the target promotion content set and send the target promotion content to the object.

[0174] In one embodiment, reference Figure 6 The detection unit 303 may include: The first acquisition subunit 3031 can be used to acquire the object’s behavior information in response to the promotional content; The first analysis subunit 3032 can be used to perform statistical analysis on the behavioral information to obtain the behavioral characteristics of the object in response to the promotional content. The detection subunit 3033 can be used to detect the historical behavioral feature data of the object in relation to the promotional content if the behavioral features meet the preset negative behavioral conditions.

[0175] In one embodiment, the first acquisition subunit 3031 can be used to: Adjust the content promotion evaluation threshold in the content promotion system; based on the adjusted content promotion evaluation threshold, send promotional content to the target; obtain the target's behavior information regarding the promotional content within a preset time interval.

[0176] In one embodiment, reference Figure 7 The first recall unit 304 may include: The preference determination subunit 3041 can be used to determine the object's preference information for the promoted content based on the historical behavioral feature data; The first recall subunit 3042 can be used to recall content based on the preference information for the set of promotional content to obtain the target set of promotional content for the object.

[0177] In one embodiment, the first recall subunit 3042 can be used to: Based on the preference information, the object group to which the object belongs is determined; promotional content associated with the object group is selected from the promotional content set; the selected promotional content is sorted to obtain sorted promotional content; based on the sorting result, content retrieval is performed on the sorted promotional content to obtain the target promotional content set for the object.

[0178] In one embodiment, the first recall subunit 3042 can be used to: Obtain the content description information of the promotional content; based on the preference information and the content description information, perform content preference matching between the object and the promotional content; based on the matching result, perform content retrieval for the set of promotional content to obtain the target set of promotional content for the object.

[0179] In one embodiment, reference Figure 8 The first recall unit 304 may include: The model identifies subunit 3043, which can be used to determine the original promotion evaluation model for the promoted content; The feature extraction subunit 3044 can be used to extract features from the historical behavior feature data to obtain the content interaction features of the object in response to the promotional content; The feature addition subunit 3045 can be used to add the content interaction feature to the original promotion evaluation model, so that the promotion evaluation model after addition can determine the content promotion evaluation result of the promotion content corresponding to the object by referring to the content interaction feature. The second recall subunit 3046 can be used to recall the set of promotional content based on the content promotion evaluation results, so as to obtain the target set of promotional content for the object.

[0180] In one embodiment, reference Figure 9 The second recall unit 305 may include: The information acquisition subunit 3051 can be used to acquire promotion statistics of the promotional content and the object's preference information for the promotional content; The parameter determination subunit 3052 can be used to determine content promotion evaluation parameters for the promoted content based on the promotion statistics and the preference information. The first sorting subunit 3053 can be used to sort the promotion content set based on the content promotion evaluation parameters in order to recall the promotion content of the object and obtain the target promotion content set of the object.

[0181] In one embodiment, the first sorting subunit 3053 can be used for: Determine the original promotion evaluation model for the promoted content; add the content promotion evaluation parameters to the original promotion evaluation model so that the added promotion evaluation model determines the content promotion evaluation result of the promoted content corresponding to the object based on the content promotion evaluation parameters; perform content sorting processing on the set of promoted content based on the content promotion evaluation result; recall the promoted content of the object based on the sorting result to obtain the target promotion content set of the object.

[0182] In one embodiment, reference Figure 10 The second recall unit 305 may include: The attribute acquisition subunit 3054 can be used to acquire the content attribute information of the promotional content and the object attribute information of the object; The second sorting subunit 3055 can be used to sort the promotional content set based on the content attribute information and the object attribute information, so as to recall the promotional content of the object and obtain the target promotional content set of the object.

[0183] In one embodiment, the second sorting subunit 3055 can be used for: Based on the content attribute information and the object attribute information, select promotional content to be sorted from the promotional content set; sort the promotional content to be sorted to obtain sorted promotional content; based on the sorting result, recall the promotional content of the object from the sorted promotional content to obtain the target promotional content set of the object.

[0184] In one embodiment, reference Figure 11 The device for processing the promotional content further includes: The feature determination unit 307 can be used to determine the target behavioral features of the object in relation to the target promotion content; The type determination unit 308 can be used to determine the object type of the object for the promotional content based on the target behavior characteristics and the preset negative behavior conditions.

[0185] In practice, each of the above units can be implemented as an independent entity or can be arbitrarily combined to be implemented as the same or several entities. For the specific implementation of each of the above units, please refer to the previous method embodiments, which will not be repeated here.

[0186] As can be seen from the above, in the promotional content processing device of this embodiment, the acquisition unit 301 acquires the promotional content request of the object; the set determination unit 302 determines the promotional content set according to the promotional content request, wherein the promotional content set includes at least one promotional content; the detection unit 303 detects the historical behavior feature data of the object towards the promotional content if the object's behavior features towards the promotional content meet the preset negative behavior conditions; the first recall unit 304 performs content recall on the promotional content set based on the historical behavior feature data when the object has the historical behavior feature data, to obtain the target promotional content set of the object; the second recall unit 305 performs content sorting processing on the promotional content set when the object does not have the historical behavior feature data, to recall the object's promotional content, to obtain the target promotional content set of the object; and the content determination unit 306 determines the target promotional content of the object from the target promotional content set and sends the target promotional content to the object.

[0187] This solution targets objects whose behavioral characteristics in response to promoted content meet preset negative behavioral conditions. After receiving a request for promoted content from a target object, it determines how to retrieve promoted content from the target object's content set based on whether the target object has historical behavioral data related to the promoted content. This allows for precise content promotion to the target object. Specifically, if the target object has historical behavioral data related to the promoted content, the solution can retrieve content based on this data. This allows for accurate prediction of the promoted content that the target object might be interested in by referring to the target object's content interaction history. If the target object does not have historical behavioral data related to the promoted content, the solution can retrieve corresponding promoted content from the promoted content set through sorting exploration, increasing the probability of predicting the promoted content that the target object might be interested in. Therefore, this solution improves the accuracy and efficiency of content promotion to the target object by increasing the processing efficiency of promoted content.

[0188] Furthermore, embodiments of this application also provide a computer device, which can be a server or terminal, etc. Figure 12 As shown, it illustrates a structural schematic diagram of the computer device involved in the embodiments of this application, specifically: The computer device may include a memory 401 having one or more computer-readable storage media, an input unit 402, a processor 403 including one or more processing cores, and a power supply 404, etc. Those skilled in the art will understand that... Figure 12The computer device structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein: The memory 401 can be used to store software programs and modules. The processor 403 executes various functional applications and data processing by running the software programs and modules stored in the memory 401. The memory 401 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device (such as audio data, telephone directory, etc.). In addition, the memory 401 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 401 may also include a memory controller to provide access to the memory 401 by the processor 403 and the input unit 402.

[0189] Input unit 402 can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical, or trackball signal inputs related to user settings and function control. Specifically, in one embodiment, input unit 402 may include a touch-sensitive surface and other input devices. A touch-sensitive surface, also known as a touch display or touchpad, can collect user touch operations on or near it (e.g., user operations using fingers, styluses, or any suitable object or accessory on or near the touch-sensitive surface) and drive corresponding connection devices according to a pre-set program. Optionally, the touch-sensitive surface may include a touch detection device and a touch controller. The touch detection device detects the user's touch orientation and the signal generated by the touch operation, transmitting the signal to the touch controller; the touch controller receives touch information from the touch detection device, converts it into touch point coordinates, sends it to the processor 403, and can receive and execute commands from the processor 403. Furthermore, various types of touch-sensitive surfaces, such as resistive, capacitive, infrared, and surface acoustic wave, can be used. In addition to the touch-sensitive surface, input unit 402 may also include other input devices. Specifically, other input devices may include, but are not limited to, one or more of the following: physical keyboard, function keys (such as volume control buttons, power buttons, etc.), trackball, mouse, joystick, etc. The processor 403 is the control center of the computer device. It connects various parts of the mobile phone via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 401, and by calling data stored in the memory 401, it performs various functions of the computer device and processes data, thereby controlling the mobile phone as a whole. Optionally, the processor 403 may include one or more processing cores; preferably, the processor 403 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 403.

[0190] The computer device also includes a power supply 404 (such as a battery) that supplies power to various components. Preferably, the power supply can be logically connected to the processor 403 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 404 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0191] Although not shown, the computer device may also include a camera, Bluetooth module, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 403 in the computer device loads the executable files corresponding to the processes of one or more applications into the memory 401 according to the following instructions, and the processor 403 runs the applications stored in the memory 401 to realize various functions, as follows: The process involves: obtaining a request for promotional content from an object; determining a set of promotional content based on the request, wherein the set of promotional content includes at least one piece of promotional content; if the object's behavioral characteristics regarding the promotional content meet preset negative behavioral conditions, detecting the object's historical behavioral characteristic data regarding the promotional content; when the object possesses the historical behavioral characteristic data, performing content recall based on the historical behavioral characteristic data for the set of promotional content to obtain the object's target set of promotional content; when the object does not possess the historical behavioral characteristic data, performing content sorting processing on the set of promotional content to recall the object's promotional content to obtain the object's target set of promotional content; determining the object's target promotional content from the target set of promotional content, and sending the target promotional content to the object.

[0192] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0193] The computer device in this embodiment can target objects whose behavioral characteristics towards promoted content meet preset negative behavioral conditions. After obtaining a promotional content request from a target object, it determines how to recall the target object's promotional content from the promotional content set based on whether the target object has historical behavioral characteristic data regarding the promotional content, thereby achieving precise content promotion targeting the target object. Specifically, if the target object has historical behavioral characteristic data regarding the promotional content, the computer device in this embodiment can recall content based on this historical behavioral characteristic data. In this way, by referring to the target object's content interaction history regarding the promotional content, it can accurately predict the promotional content that the target object is interested in. If the target object does not have historical behavioral characteristic data regarding the promotional content, the computer device in this embodiment can recall the corresponding promotional content from the promotional content set through sorting exploration, thereby increasing the probability of predicting the promotional content that the target object is interested in. Therefore, when promoting content to a target object, the computer device in this embodiment can improve the accuracy and efficiency of promotion by improving the processing efficiency of promotional content.

[0194] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0195] Therefore, embodiments of this application provide a storage medium storing a plurality of instructions that can be loaded by a processor to execute steps in any of the promotional content processing methods provided in embodiments of this application. For example, the instructions can execute the following steps: The process involves: obtaining a request for promotional content from an object; determining a set of promotional content based on the request, wherein the set of promotional content includes at least one piece of promotional content; if the object's behavioral characteristics regarding the promotional content meet preset negative behavioral conditions, detecting the object's historical behavioral characteristic data regarding the promotional content; when the object possesses the historical behavioral characteristic data, performing content recall based on the historical behavioral characteristic data for the set of promotional content to obtain the object's target set of promotional content; when the object does not possess the historical behavioral characteristic data, performing content sorting processing on the set of promotional content to recall the object's promotional content to obtain the object's target set of promotional content; determining the object's target promotional content from the target set of promotional content, and sending the target promotional content to the object.

[0196] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0197] The storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0198] Since the instructions stored in the storage medium can execute the steps in any of the promotional content processing methods provided in the embodiments of this application, the beneficial effects that any of the promotional content processing methods provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.

[0199] According to one 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 methods provided in various alternative implementations of the processing aspects of the above-described promotional content.

[0200] The foregoing has provided a detailed description of a method, apparatus, computer device, and storage medium for processing promotional content provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for processing promotional content, characterized in that, include: Request the promotional content of the object; A set of promotional content is determined based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content; If the object's behavior characteristics towards the promoted content meet preset negative behavior conditions, then the historical behavior characteristic data of the object towards the promoted content is detected, including: lowering the content promotion evaluation threshold in the content promotion system; sending promoted content to the object based on the lowered content promotion evaluation threshold to increase the object's exposure of the promoted content; obtaining the object's behavior information towards the promoted content within a preset time interval; performing statistical analysis on the behavior information to obtain the object's behavior characteristics towards the promoted content; if the behavior characteristics meet preset negative behavior conditions, then the object is determined to be a silent user, and the historical behavior characteristic data of the object towards the promoted content is detected; wherein, the preset negative behavior conditions include: within the preset time interval, the object's exposure of the promoted content exceeds a preset exposure threshold, and the object's reaction to the promoted content is a negative reaction; When the object has the historical behavior feature data, content recall is performed on the set of promotional content based on the historical behavior feature data to obtain the target set of promotional content for the object. When the object does not have the historical behavior feature data, the promotion content set is sorted to recall the object's promotion content and obtain the object's target promotion content set. The target promotion content for the object is determined from the set of target promotion content, and the target promotion content is sent to the object; Determine the target behavioral characteristics of the object in relation to the target promotional content; Based on the target behavioral characteristics and the preset negative behavioral conditions, the target type of the object is determined to be related to the promotional content; if the target behavioral characteristics meet the preset negative behavioral conditions, the object is determined to be a type of object that is indifferent to the promotional content, and the exposure of the object to the promotional content is reduced; if the target behavioral characteristics do not meet the preset negative behavioral conditions, the object is determined to be a type of object that is responsive to the promotional content, so as to convert the silent user into a regular user.

2. The method for processing promotional content according to claim 1, characterized in that, Based on the historical behavioral feature data, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the object, including: Based on the historical behavioral characteristic data, determine the object's preference information for the promoted content; Based on the preference information, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the object.

3. The method for processing promotional content according to claim 2, characterized in that, Based on the preference information, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the object, including: Based on the preference information, the object group to which the object belongs is determined; Select promotional content associated with the target group from the set of promotional content; The selected promotional content is sorted to obtain the sorted promotional content; Based on the ranking results, content retrieval is performed on the ranked promotional content to obtain the target promotional content set for the object.

4. The method for processing promotional content according to claim 2, characterized in that, Based on the preference information, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the object, including: Obtain the content description information of the promotional content; Based on the preference information and the content description information, the object and the promotional content are matched according to their content preferences. Based on the matching results, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the object.

5. The method for processing promotional content according to claim 1, characterized in that, Based on the historical behavioral feature data, content retrieval is performed on the set of promotional content to obtain the target set of promotional content for the object, including: Determine the original promotion evaluation model for the promoted content; Feature extraction is performed on the historical behavioral feature data to obtain the content interaction features of the object in response to the promotional content; The content interaction features are added to the original promotion evaluation model so that the promotion evaluation model, after adding the content interaction features, can determine the content promotion evaluation result of the promotion content corresponding to the object by referring to the content interaction features. Based on the content promotion evaluation results, content recall is performed on the set of promotional content to obtain the target set of promotional content for the object.

6. The method for processing promotional content according to claim 1, characterized in that, The promotional content set is sorted to retrieve the promotional content of the object, resulting in a target promotional content set for the object, including: Obtain promotion statistics of the promoted content and the object's preference information for the promoted content; Based on the promotion statistics and the preference information, determine the content promotion evaluation parameters for the promoted content; Based on the content promotion evaluation parameters, the promotion content set is sorted to recall the promotion content of the object, thereby obtaining the target promotion content set of the object.

7. The method for processing promotional content according to claim 6, characterized in that, Based on the content promotion evaluation parameters, the promotion content set is sorted to recall the promotion content of the object, resulting in the target promotion content set of the object, including: Determine the original promotion evaluation model for the promoted content; The content promotion evaluation parameters are added to the original promotion evaluation model so that the promotion evaluation model, after adding the parameters, determines the content promotion evaluation result of the promoted content corresponding to the object. Based on the content promotion evaluation results, the promotion content set is sorted. Based on the ranking results, the promotional content of the object is recalled to obtain the target promotional content set of the object.

8. The method for processing promotional content according to claim 1, characterized in that, The promotional content set is sorted to retrieve the promotional content of the object, resulting in a target promotional content set for the object, including: Obtain the content attribute information of the promotional content and the object attribute information of the object; Based on the content attribute information and the object attribute information, the promotion content set is sorted to recall the promotion content of the object, thereby obtaining the target promotion content set of the object.

9. The method for processing promotional content according to claim 8, characterized in that, Based on the content attribute information and the object attribute information, the promotional content set is sorted to recall the promotional content of the object, resulting in a target promotional content set for the object, including: Based on the content attribute information and the object attribute information, select promotional content to be sorted from the promotional content set; The promotional content to be sorted is sorted to obtain the sorted promotional content; Based on the sorting results, the promotional content of the object is recalled from the sorted promotional content to obtain the target promotional content set of the object.

10. A device for processing promotional content, characterized in that, include: The retrieval unit is used to retrieve the promotional content request for an object. A set determination unit is used to determine a set of promotional content based on the promotional content request, wherein the set of promotional content includes at least one piece of promotional content; The detection unit is configured to detect the historical behavioral characteristic data of the object towards the promoted content if the object's behavioral characteristics towards the promoted content meet a preset negative behavioral condition. This includes: lowering the content promotion evaluation threshold in the content promotion system; sending promoted content to the object based on the lowered content promotion evaluation threshold to increase the object's exposure to the promoted content; acquiring the object's behavioral information towards the promoted content within a preset time interval; performing statistical analysis on the behavioral information to obtain the object's behavioral characteristics towards the promoted content; and if the behavioral characteristics meet the preset negative behavioral condition, determining the object as a silent user and detecting the object's historical behavioral characteristic data towards the promoted content. The preset negative behavioral condition includes: within the preset time interval, the object's exposure to the promoted content exceeds a preset exposure threshold, and the object's reaction to the promoted content is a negative reaction. The first recall unit is used to recall content based on the historical behavior feature data of the object when the object has the historical behavior feature data, so as to obtain the target promotion content set of the object. The second recall unit is used to sort the promotional content set when the object does not have the historical behavior feature data, so as to recall the object's promotional content and obtain the object's target promotional content set. The content determination unit is used to determine the target promotion content of the object from the target promotion content set, and send the target promotion content to the object; A feature determination unit is used to determine the target behavioral features of the object in relation to the target promotional content; The type determination unit is used to determine the object type of the object in relation to the promotional content based on the target behavioral characteristics and the preset negative behavioral conditions. If the target behavioral characteristics meet the preset negative behavioral conditions, the object is determined to be an object type that is indifferent to the promotional content, and the exposure of the object to the promotional content is reduced. If the target behavioral characteristics do not meet the preset negative behavioral conditions, the object is determined to be an object type that is responsive to the promotional content, so as to convert the silent user into a regular user.

11. An electronic device, characterized in that, It includes a memory and a processor; the memory stores an application program, and the processor runs the application program within the memory to perform the operations in the method for processing promotional content as described in any one of claims 1 to 9.

12. A storage medium, characterized in that, The storage medium stores a plurality of instructions adapted for loading by a processor to execute the steps of the method for processing promotional content as described in any one of claims 1 to 9.

13. A computer program product, characterized in that, It includes computer instructions, which, when executed, implement the steps of the method for processing promotional content as described in any one of claims 1 to 9.

Images