An information promotion method, device, equipment and computer readable storage medium

Abstract

The application provides an information promotion method and device, equipment and a computer readable storage medium; relates to information processing technology and blockchain technology; the method comprises the following steps: obtaining the historical operation data of a target object; determining the object group to which the target object belongs according to the historical operation data, wherein the object group comprises objects having the same operation mode for promotion information; optimizing the information queue of the target object based on the optimization strategy corresponding to the object group to obtain an optimized queue, wherein the optimized queue comprises candidate promotion information selected for the target object; selecting target promotion information from the optimized queue and pushing the target promotion information to the target object. Through the application, the pertinence of information promotion can be improved.

Description

Technical Field

[0001] This application relates to information processing technology and blockchain technology, and in particular to an information dissemination method, apparatus, device and computer-readable storage medium. Background Technology

[0002] Information promotion aims to maximize the exposure of information to promote related products, news, etc., to users. Promotional information typically goes through stages such as recall, initial ranking, final ranking, online strategy screening, and information feedback before being recommended to users.

[0003] In related technologies, information promotion systems typically target all users and use the same strategies to filter promotional information. However, the same strategy may not be suitable for all types of users. For example, different types of users have different tolerance levels for promotional information and different conversion rates. Therefore, information promotion in related technologies lacks specificity. Summary of the Invention

[0004] This application provides an information promotion method, apparatus, device, and computer-readable storage medium, which can improve the targeting of information promotion.

[0005] The technical solution of this application embodiment is implemented as follows: This application provides an information promotion method, including: Retrieve historical operation data of the target object; Based on the historical operation data, the target object is identified to which it belongs, wherein the target object group includes objects that have the same operation pattern in response to promotional information; Based on the optimization strategy corresponding to the target group, the information queue of the target object is optimized to obtain an optimized queue, which includes candidate promotion information selected for the target object; Target promotion information is selected from the optimization queue and pushed to the target object.

[0006] This application provides an information promotion device, including: The data acquisition module is used to acquire historical operation data of the target object; The group determination module is used to determine the target group to which the target object belongs based on the historical operation data, wherein the target group includes objects that have the same operation mode for the promotional information; The queue optimization module is used to optimize the information queue of the target object based on the optimization strategy corresponding to the target group, so as to obtain an optimized queue, wherein the optimized queue includes candidate promotion information selected for the target object; The information push module is used to filter out target promotion information from the optimization queue and push the target promotion information to the target object.

[0007] In some embodiments of this application, the target group includes: objects whose exposure count exceeds a first exposure threshold; the queue optimization module is further configured to select replacement promotional information from various promotional information in the platform information pool; and use the replacement promotional information to replace the promotional information at the tail of the fine-ranking queue in the information queue of the target object to obtain the optimized queue.

[0008] In some embodiments of this application, the target group includes: objects whose exposure value exceeds a first value threshold; the queue optimization module is further configured to extend the length of the coarse-order queue in the information queue of the target object to obtain an expanded queue; the number of promotional information in the expanded queue is greater than the number of promotional information in the coarse-order queue; and the promotional information in the expanded queue is fine-ordered to obtain the optimized queue.

[0009] In some embodiments of this application, the target group includes: objects whose exposure value is equal to or lower than a second value threshold; the queue optimization module is further configured to shorten the length of the coarse queue in the information queue of the target object to obtain a simplified queue; the number of promotional information in the simplified queue is less than the number of promotional information in the coarse queue; and the promotional information in the simplified queue is finely sorted to obtain the optimized queue.

[0010] In some embodiments of this application, the target group includes: objects with a click-through rate lower than a click-through rate threshold; the queue optimization module is further configured to acquire historical search data and historical conversion data of the target objects; use the historical search data and the historical conversion data to explore the interests of the target objects and obtain interest tags; select interest promotion information from the platform information pool based on the interest tags; and optimize the information queue using the interest promotion information to obtain the optimized queue.

[0011] In some embodiments of this application, the queue optimization module is further configured to replace the promotional information in the information queue with the interest promotional information to obtain the optimized queue; or, to add the interest promotional information to the information queue to obtain the optimized queue.

[0012] In some embodiments of this application, the queue optimization module is further configured to determine the current industry category to which the promotional information in the information queue of the target object belongs; select other promotional information from the platform information pool that is different from the current industry category; and add the other promotional information to the information queue to obtain the optimized queue.

[0013] In some embodiments of this application, the object group includes: objects whose exposure count is less than or equal to a second exposure threshold, and objects whose exposure value is less than or equal to a second value threshold; The information promotion device also includes: a frequency adjustment module; The frequency adjustment module is used to reduce the frequency of information promotion to the target object.

[0014] In some embodiments of this application, the historical operation data includes the number of exposures, exposure value, and click-through rate corresponding to the target object; the group determination module is further configured to compare the number of exposures of the target object with a first exposure threshold and a second exposure threshold respectively to determine the object group to which the target object belongs; or, compare the exposure value of the target object with a first value threshold and a second value threshold respectively to determine the object group to which the target object belongs; or, compare the click-through rate of the target object with a click-through rate threshold to determine the object group to which the target object belongs.

[0015] In some embodiments of this application, the information promotion device further includes: a threshold determination module; the threshold determination module is used to analyze the average daily exposure of objects on the information platform to obtain a first exposure threshold and a second exposure threshold; the average daily exposure represents the number of promotional information viewed by objects on the information platform per day; analyze the exposure value of promotional information in the platform information pool to obtain a first value threshold and a second value threshold; and analyze the click-through rate of promotional information in the platform information pool to obtain the click-through rate threshold.

[0016] In some embodiments of this application, the threshold determination module is further configured to arrange the exposure value of the promotional information in the platform information pool from largest to smallest to obtain an exposure value sequence; determine the first exposure value in the exposure value sequence as the first value threshold, and determine the second exposure value in the exposure value sequence as the second value threshold.

[0017] In some embodiments of this application, the threshold determination module is further configured to: filter out promotional information to be analyzed from the promotional information in the platform information pool that has been displayed more than a preset number of times; arrange the click-through rates of the promotional information to be analyzed in order of magnitude to obtain a click-through rate sequence; and determine the click-through rate of the preset position in the click-through rate sequence as the click-through rate threshold.

[0018] This application provides an information promotion device, including: Memory, used to store executable information and instructions; The processor, when executing executable information promotion instructions stored in the memory, implements the information promotion method provided in the embodiments of this application.

[0019] This application provides a computer-readable storage medium storing executable information promotion instructions, which, when executed by a processor, implement the information promotion method provided in this application.

[0020] The embodiments of this application have the following beneficial effects: The information promotion device utilizes the historical operation data of the target object to determine the target group to which the target object belongs. Then, based on the optimization strategy corresponding to the target group, it optimizes the information queue of the target object to obtain an optimized queue corresponding to the target object. Next, it filters out target promotion information from the optimized queue and promotes it. In this way, the information promotion process of the target object can be optimized according to its characteristics, so that the information promotion methods differ between different objects, thus improving the targeting of information promotion. Attached Figure Description

[0021] Figure 1A This is an optional architecture diagram of the information promotion system provided in the embodiments of this application; Figure 1B This is another optional architecture diagram of the information promotion system provided in the embodiments of this application; Figure 2 This is a schematic diagram of the structure of the information promotion device provided in the embodiments of this application; Figure 3 This is an optional flowchart illustrating the information promotion method provided in this application embodiment; Figure 4 This is a schematic diagram illustrating the optimization of the queue by replacing promotional information, provided in an embodiment of this application. Figure 5 This is a schematic diagram illustrating the optimization of a queue by extending the length of a coarse-sorted queue, provided in an embodiment of this application. Figure 6 This is a schematic diagram illustrating the optimization of a queue by shortening the length of a coarse-sorted queue, provided in an embodiment of this application. Figure 7 This is an optional flowchart illustrating the information promotion method provided in the embodiments of this application. Figure 2 ; Figure 8 This is a schematic diagram illustrating how to obtain an optimized queue through other promotional information, as provided in an embodiment of this application. Figure 9 This is a user percentage distribution chart for statistical CPM segments provided in the embodiments of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0024] In the following description, the terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first" and "second" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0026] In the implementation of this application, the collection and processing of relevant data should strictly comply with the requirements of relevant laws and regulations, obtain the informed consent or separate consent of the personal information subject, and carry out subsequent data use and processing within the scope of laws and regulations and the authorization of the personal information subject.

[0027] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0028] 1) Cloud technology refers to a managed technology that unifies a series of resources such as hardware, software, and networks within a wide area network or local area network to realize the computing, storage, processing, and sharing of data.

[0029] Cloud technology is a general term encompassing network technology, information technology, integration technology, management platform technology, and application technology based on the cloud computing business model. It can form resource pools, providing flexible and convenient on-demand access. Cloud computing technology will become a crucial support. Backend services of technical network systems require substantial computing and storage resources, such as video websites, image websites, and many portal websites. With the rapid development and application of the internet industry, every item may have its own identification mark in the future, requiring data to be transmitted to the backend for logical processing. Data at different levels will be processed separately, and various industry data will all require robust system support, which can only be achieved through cloud technology.

[0030] 2) Promotional information refers to information used for publicity and promotion, such as product order links and promotional introductions of high-quality information. Promotional information is generally placed on information platforms by the information creators, i.e., advertisers, and then recommended to users by the information platforms.

[0031] 3) Exposure count refers to the total number of times a user views the promotional information.

[0032] 4) Click count is the total number of times a user clicks on promotional information.

[0033] 5) CPM (Cost Per Mile), which is the revenue generated from each impression of a promotional message / impressions * 1000. CPM is a comprehensive reflection of the value of traffic.

[0034] 6) CTR (Click-Through-Rate) is calculated as: number of clicks on the promotional message / number of impressions of the promotional message.

[0035] 7) eCPM (Estimated revenue per thousand impressions, effective CPM) is an important reference indicator in information promotion. For information platforms, information promotion should aim to maximize eCPM. The calculation method for eCPM is shown in equation (1): (1) in, This represents the estimated conversion rate, which is the percentage of the total number of clicks on the promotional message that the estimated number of users who reach the target number of the promotional message. This represents the advertiser's expected value cost, such as the advertiser's bid; This indicates the estimated click-through rate; In order to adjust the value cost factor (due to) (Correcting inaccuracies).

[0036] 8) The base price strategy is to remove promotional information with a price lower than the base price of the promotional slot when selecting promotional information for the promotional slot (i.e., the operation of placing promotional information in the information flow), and retain promotional information with a price higher than the base price of the promotional slot.

[0037] 9) Freshness strategy: For the same user, in the same scenario or on the same site, the same promotional information cannot be repeated within 10 minutes, and the same product promotional information cannot be repeated within 1 hour.

[0038] 10) Information feed is the display format of information retrieved by the user terminal. The information feed contains numerous slots, which are used to display different types of information, such as news and promotional information.

[0039] Information promotion aims to maximize the exposure of information to promote related products, news, etc., to users. Promotional information typically goes through stages such as recall, initial ranking, final ranking, online strategy screening, and information return before being recommended to users. Online strategy screening involves using strategies such as low-price and novelty strategies to filter promotional information, making the returned information more likely to be clicked by users and lead to conversions (e.g., forwarding, placing orders).

[0040] In related technologies, information promotion systems typically target all users, using the same strategies to filter promotional information. For example, when using a freshness strategy, all users are subject to restrictions on encountering the same promotional information within 10 minutes or the same product within one hour. However, the same strategy is not necessarily suitable for all types of users. For instance, different types of users have different tolerance levels for promotional information and different conversion rates. Therefore, information promotion in related technologies lacks specificity.

[0041] Furthermore, a lack of targeted information promotion can lead to an inability to allocate the amount of promotional information reasonably to different users, resulting in low overall platform value and poor user experience. For example, allocating too much promotional information to users with low tolerance for promotional information may cause annoyance, while allocating too little promotional information to users who like to click on promotional information and convert, thus reducing the platform's value and revenue.

[0042] To address the aforementioned issues, the applicant conducted an in-depth analysis of the current information promotion system and discovered that the coarse-grained model, when estimating conversion rates, often yields higher estimates for low-conversion-rate groups (i.e., estimated conversion rates are higher than actual conversion rates) and lower estimates for high-conversion-rate groups (i.e., estimated conversion rates are lower than actual conversion rates). This indicates that the promotional information selected by the current system lacks targeting, resulting in low overall platform revenue. To improve the platform's overall revenue, the applicant found that segmenting users and employing different strategies for information promotion tailored to different user groups can enhance the targeting of information promotion.

[0043] Based on the above ideas, embodiments of this application provide an information promotion method, apparatus, device, and computer-readable storage medium, which can improve the targeting of information promotion. The following describes exemplary applications of the information promotion device provided in this application. The information promotion device provided in this application can be implemented as various types of terminals such as laptops, tablets, desktop computers, set-top boxes, mobile devices (e.g., mobile phones, portable music players, personal digital assistants, dedicated messaging devices, portable gaming devices), in-vehicle terminals, and smart TVs. It can also be implemented as a server, or as a device cluster composed of terminals and servers. The following will describe exemplary applications when the information promotion device is implemented as a server.

[0044] See Figure 1A , Figure 1A This is an optional architecture diagram of the information promotion system provided in the embodiments of this application. In order to support an information promotion application, in the information promotion system 100, the terminal 400 connects to the server 200 through the network 300. The network 300 can be a wide area network or a local area network, or a combination of the two.

[0045] Terminal 400 is used to obtain the target object's historical operation data based on the target object's operation on the information promotion interface of the graphical interface 400-1, and send the historical operation data to server 200 through network 300. Server 200 stores the historical operation data in the database.

[0046] During information promotion, server 200 obtains historical operation data of the target object; based on the historical operation data, it determines the target object group to which the target object belongs, wherein the target object group includes objects with the same operation mode for the promotion information; based on the optimization strategy corresponding to the target object group, it optimizes the information queue of the target object to obtain an optimized queue, which includes candidate promotion information selected for the target object; it selects the target promotion information from the optimized queue and pushes the target promotion information to the target object, that is, it sends the target promotion information to the terminal 400 through network 300.

[0047] Terminal 400 is also used to display targeted promotional information on the graphical interface 400-1.

[0048] In some embodiments, server 200 may be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. Terminal 400 may be a smartphone, tablet computer, laptop computer, desktop computer, smart speaker, smartwatch, in-vehicle terminal, smart TV, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment of the invention.

[0049] This application's embodiments can also be implemented using blockchain technology; see [link to relevant documentation]. Figure 1B , Figure 1B This is another optional architecture diagram of the information promotion system provided in the embodiments of this application. Figure 1B In the information promotion system 100, the server 200 uploads the target promotion information determined for the target object to the blockchain to ensure that the target promotion information can be recorded and that the record cannot be tampered with.

[0050] In some embodiments, server 200 can join blockchain network 700 and become a node within it. The type of blockchain network 700 is flexible and diverse; for example, it can be any of a public blockchain, private blockchain, or consortium blockchain. Taking a public blockchain as an example, any electronic device of any business entity can access blockchain network 700 without authorization to act as a consensus node in blockchain network 700. For example, terminal 600 can be mapped to consensus node 700-1 in blockchain network 700 to achieve consensus when uploading metadata to the blockchain.

[0051] Taking blockchain network 700 as a consortium blockchain as an example, terminal 600 can connect to blockchain network 700 and become a node after obtaining authorization. Server 200 obtains the historical operation data of the target object, determines the target object's target group based on the historical operation data, optimizes the target object's information queue based on the optimization strategy corresponding to the target group, obtains an optimized queue, filters out target promotion information from the optimized queue, and sends the target object's identity identifier and target promotion information to blockchain network 700 for consensus. When consensus is passed, blockchain network 700 allocates a new block for the target object's identity identifier and target promotion information, writes the target object's identity identifier and target promotion information into the new block, and adds the new block to the end of the blockchain.

[0052] See Figure 2 , Figure 2 This is a schematic diagram of the structure of the information promotion device provided in the embodiments of this application. Figure 2 The information dissemination device 500 shown includes at least one processor 510, a memory 550, at least one network interface 520, and a user interface 530. The various components in the information dissemination device 500 are coupled together via a bus system 540. It is understood that the bus system 540 is used to implement communication between these components. In addition to a data bus, the bus system 540 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in… Figure 2 The general labeled all buses as Bus System 540.

[0053] The processor 510 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0054] User interface 530 includes one or more output devices 531 that enable the presentation of media content, including one or more speakers and / or one or more visual displays. User interface 530 also includes one or more input devices 532, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0055] The memory 550 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state storage, hard disk drives, optical disk drives, etc. The memory 550 may optionally include one or more storage devices physically located away from the processor 510.

[0056] The memory 550 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), and the volatile memory may be random access memory (RAM). The memory 550 described in this application embodiment is intended to include any suitable type of memory.

[0057] In some embodiments, memory 550 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, as illustrated below.

[0058] Operating system 551 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks; The network communication module 552 is used to reach other computing devices via one or more (wired or wireless) network interfaces 520, exemplary network interfaces 520 including: Bluetooth, Wi-Fi, and Universal Serial Bus (USB), etc. Presentation module 553 is used to enable the presentation of information (e.g., user interface for operating peripheral devices and displaying content and information) via one or more output devices 531 (e.g., display screen, speaker, etc.) associated with user interface 530. The input processing module 554 is used to detect and translate one or more user inputs or interactions from one or more input devices 532. In some embodiments, the apparatus provided in this application can be implemented in software. Figure 2 An information promotion device 555 stored in memory 550 is shown. It can be software in the form of programs and plug-ins, including the following software modules: data acquisition module 5551, group determination module 5552, queue optimization module 5553, information push module 5554, frequency adjustment module 5555, and threshold determination module 5556. These modules are logically related and can therefore be arbitrarily combined or further split according to the functions they implement.

[0059] The functions of each module will be explained below.

[0060] In other embodiments, the information promotion device provided in this application can be implemented in hardware. As an example, the information promotion device provided in this application can be a processor in the form of a hardware decoding processor, which is programmed to execute the information promotion method provided in this application. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0061] For example, an embodiment of this application provides an information promotion device, including: Memory, used to store executable information and instructions; The processor, when executing executable information promotion instructions stored in the memory, implements the information promotion method provided in the embodiments of this application.

[0062] The following will describe the information promotion method provided in this application embodiment, with reference to exemplary applications and implementations of the information promotion device provided in the embodiments of this application. It should be noted that when the information promotion device is implemented as a device cluster composed of terminals and servers, the information promotion method provided in the embodiments of this application can be implemented through cloud technology.

[0063] See Figure 3 , Figure 3 This is an optional flowchart of the information promotion method provided in the embodiments of this application, which will be combined with... Figure 3 The steps shown are explained.

[0064] S101. Obtain historical operation data of the target object.

[0065] This application embodiment is implemented in a scenario of information promotion targeting a specific object, such as pushing promotional videos, text descriptions, or order links for products awaiting promotion to the target object. When information promotion begins, the information promotion device first retrieves and acquires the target object's historical operation data to analyze the target object and achieve information promotion.

[0066] Understandably, information promotion devices can begin the information promotion process targeting a specific audience at a preset time, such as acquiring the audience's historical activity data at 12:00 noon each day for information promotion. Alternatively, the device can initiate the information promotion process when the audience requests new information, such as acquiring their historical activity data when they refresh a news page for information promotion.

[0067] It should be noted that historical operation data can be the target object's operation data on promotional information within a preset historical time period, or it can be all operation data of the target object from registration to the current moment.

[0068] Furthermore, the preset historical time period can be the 30 days prior to the current time or the 12 hours prior to the current time; this application does not impose any limitation on this.

[0069] Operational data for promotional information may include, but is not limited to, login count, refresh count, click-through rate, browsing duration, collection count, like count, forward count, comment count, question count, order count, add to cart count, exposure count (i.e., the number of pieces of information viewed by the target audience), exposure value (e.g., value per thousand impressions), etc., which are not limited in this application.

[0070] It is understandable that the target is any user registered on the information platform.

[0071] S102. Determine the target object group based on historical operation data.

[0072] Historical operational data reflects the different actions taken by target users in response to promotional information, and these actions can characterize the target users' preferences and attitudes towards the promotional information. Therefore, information promotion equipment analyzes the historical operational data of target users to extract their operational patterns towards promotional information, and then identifies groups with the same operational patterns as the target users, thus obtaining the target user's target group. In other words, the target group includes individuals with similar operational patterns towards promotional information, such as those who like to click on promotional information or those who like to refresh the page immediately after browsing a small amount of information.

[0073] In some embodiments, the information promotion device may use the number of exposures to determine the target group, or it may use the click rate to determine the target group, or it may use the exposure value to determine the target group. This application does not limit the scope of the invention.

[0074] In other embodiments, the information promotion device can also classify target objects by combining exposure quantity, click quantity, and exposure value simultaneously, thereby classifying target objects into objects whose exposure value exceeds a first value threshold, objects whose exposure quantity exceeds a first exposure threshold, objects whose exposure value is equal to or lower than a second value threshold, objects whose click quantity is lower than a click-through rate threshold, but whose exposure value exceeds a first value threshold, and so on.

[0075] S103. Based on the optimization strategy corresponding to the object group, optimize the information queue of the target object to obtain the optimized queue.

[0076] After identifying the target audience, the information promotion equipment will select the optimal strategy corresponding to the target audience from the preset strategies for different groups. Then, it will use the optimization strategy to optimize the information queue of the target audience. The optimized information queue is called the optimized queue.

[0077] It should be noted that the information queue includes the initial promotional information selected for the target audience. The information queue may include the coarse-ranked queue determined by the coarse-ranking model, and may also include the fine-ranked queue determined by the fine-ranking model; this application does not limit this. Furthermore, the recall model, coarse-ranking model, and fine-ranking model together constitute the information promotion system.

[0078] It is understandable that the promotional information in the information queue reflects the existing information promotion process targeting the target audience (i.e., the corresponding information queue can only be determined through the existing information promotion process). Therefore, the information queue is optimized using the optimization strategies for the target audience's group. Essentially, this means using different strategies to optimize the information promotion process for different audiences. For example, for audiences whose exposure value is equal to or equal to the second threshold (these audiences tend to order low-value products), the computing power allocation during information promotion is reduced to utilize less computing power (computing resources, such as memory size and number of cores) to complete the information promotion for the target audience (since these audiences contribute less value to the information platform, more computing power needs to be provided to audiences with higher value contributions to improve the overall value of the platform). For audiences whose exposure count is lower than or equal to the second exposure threshold, and whose exposure value is lower than or equal to the second value threshold, the frequency of candidate promotional information is reduced to avoid generating negative feelings among the target audience.

[0079] Information promotion equipment can obtain an optimized queue by directly modifying the fine-ranked queue in the information queue, or it can modify the coarse-ranked queue in the information queue and then perform fine-ranking on the modified coarse-ranked queue to obtain an optimized queue.

[0080] The optimization queue includes candidate promotional information selected for the target audience, and the promotional information pushed to the target audience is high-quality information filtered from the candidate promotional information in the information queue. In other words, the optimization queue is the information queue that directly receives the target promotional information.

[0081] S104. Filter out target promotion information from the optimization queue and push the target promotion information to the target object.

[0082] The information promotion device selects the N promotional messages with the highest comprehensive index ranking from the optimization queue as target promotional messages, or it can arbitrarily select N promotional messages from the optimization queue as target promotional messages. Then, it pushes the target promotional messages to the terminals used by the target audience through the network, thereby completing the information promotion process. Here, N is greater than or equal to 1.

[0083] In some embodiments, the overall metric can be eCPM (estimated revenue per thousand impressions). In this case, the information promotion device can calculate the eCPM for each promotion message in the optimization queue, and then use the N promotion messages with the highest eCPM, such as the single promotion message with the highest eCPM, or the three promotion messages with the highest eCPM, as the target promotion messages.

[0084] In other embodiments, the comprehensive indicator can also be pCTR (predicted click-through rate). In this way, the information promotion device can calculate the pCTR for each promotion message in the optimization queue and select the N promotion messages with the largest pCTR as the target promotion messages.

[0085] In this embodiment, the information promotion device utilizes the historical operation data of the target object to determine the target group to which the target object belongs. Then, based on the optimization strategy corresponding to the target group, it optimizes the information queue of the target object to obtain an optimized queue corresponding to the target object. Next, it filters and promotes target promotion information from the optimized queue. In this way, the information promotion process can be optimized according to the characteristics of the target object, so that the information promotion methods differ between different objects, improving the targeting of information promotion. Furthermore, after improving the targeting of information promotion, the quantity and type of promotional information can be reasonably allocated to different users on a macro level. For example, the types and quantity of information pushed can be increased for users who like to click on promotional information and convert, while computing power can be reduced for users who are not interested in promotional information, so that computing power is concentrated on high-quality users, thereby improving the overall value of the information platform.

[0086] In some embodiments of this application, the object group includes objects whose exposure count exceeds a first exposure threshold. In this case, the information queue of the target objects is optimized to obtain an optimized queue. The specific implementation process of S103 may include: S1031-S1032, as follows: S1031. Select replacement promotional information from various promotional information in the platform's information pool.

[0087] S1032. By replacing the promotional information, the promotional information at the tail of the fine-sorted queue in the information queue of the target object is replaced to obtain the optimized queue.

[0088] Targets whose exposure exceeds the first exposure threshold are highly active on the platform and frequently view promotional information. Promotional information at the end of the refined ranking queue for these targets is generally poorly performing during refined ranking for various reasons. Therefore, the promotional device can determine that this information is unsuitable for the target audience and may never be pushed to them. To address this, the promotional device first randomly selects a preset number of promotional messages from the platform's information pool as replacements, or selects promotional messages similar to those clicked by the target audience. Next, the promotional device replaces the promotional messages at the end of the refined ranking queue with these replacement messages, allowing for the introduction of new promotional information during the promotional message return phase.

[0089] For example, Figure 4 This is a schematic diagram of obtaining an optimized queue by replacing promotional information according to an embodiment of this application. Queue 4-1 is the fine-ranked queue in the information queue. The information promotion device replaces the promotional information 4-11 at the end of queue 4-1 with the replacement promotional information 4-2 selected from the platform information pool. The resulting queue 4-3 is the optimized queue.

[0090] In this application embodiment, the number of exposures may refer to the number of promotional information viewed by the target object on average per day, or it may refer to the total number of promotional information viewed by the target object in historical time. This application is not limited to this.

[0091] In this embodiment, the information promotion device uses selected replacement promotion information to replace the poorly performing promotion information at the end of the fine-ranking queue. By eliminating the poorly performing promotion information and introducing new promotion information, it can provide diverse promotion information for highly active targets, thereby optimizing the information promotion process for highly active targets.

[0092] In some embodiments of this application, the target group includes objects whose exposure value exceeds a first value threshold. In this case, the information queue of the target objects is optimized to obtain an optimized queue. The specific implementation process of S103 may include: S1033-S1034, as follows: S1033. Extend the length of the coarse queue in the information queue of the target object to obtain the extended queue.

[0093] When the exposure value of an object within a target group exceeds a first value threshold, it indicates that these objects are likely to convert and are considered high-conversion targets. Furthermore, since fine-ranking typically selects promotional information with the highest exposure value to return to the target audience, the information promotion device will select some promotional information from the platform's information pool and add it to the coarse-ranking queue, thus lengthening the coarse-ranking queue and creating an expanded queue. This allows more promotional information to be sent to the fine-ranking model, increasing the number of promotional information participating in fine-ranking and thereby increasing the computing power for high-conversion targets. This increases the likelihood of pushing better-performing promotional information to the target audience (better-performing promotional information might be mistakenly displayed during the coarse-ranking and recall stages).

[0094] It is understandable that the number of promotional messages in the expanded queue is greater than the number of promotional messages in the coarse queue.

[0095] It should be noted that the information promotion device can arbitrarily select promotional information from the platform's information pool to add to the coarse-rank queue. It can also select promotional information from the platform's information pool whose exposure value is higher than the value corresponding to the conversion behavior of the target object in the historical period and add it to the coarse-rank queue. This application does not impose any restrictions on this.

[0096] It is understandable that exposure value can refer to the value per thousand exposures or the forwarding rate per thousand exposures, etc., and this application does not limit it here.

[0097] S1034. The promotion information in the expanded queue is sorted to obtain the optimized queue.

[0098] After the information promotion equipment obtains the expanded queue, it will input the promotion information in the expanded queue into the fine ranking model to evaluate and filter this promotion information, thereby obtaining the optimized queue.

[0099] For example, Figure 5 This is a schematic diagram illustrating the optimization of a queue by extending the length of a coarse-sorted queue, as provided in an embodiment of this application. Queue 5-1 is the coarse-sorted queue in the information queue. Originally, the length of queue 5-1 is 5. The information promotion device adds promotional information 5-2 and promotional information 5-3 to queue 5-1, thus obtaining an expanded queue 5-4. Then, the expanded queue 5-4 is fine-sorted to further filter the promotional information. The resulting fine-sorted queue is the optimized queue 5-5. Figure 5 As can be seen, the additional promotional information 5-2 added by the information promotion device has the opportunity to enter the fine ranking queue and thus be returned to the target audience.

[0100] In this embodiment, when the target group includes high-value objects whose exposure value exceeds a first value threshold, the information promotion device extends the coarse queue to improve the computing power for high-conversion objects, thereby increasing the possibility of pushing better-performing promotional information to the target objects and optimizing the information promotion process for high-conversion objects.

[0101] In some embodiments of this application, the object group includes objects whose exposure value is equal to or lower than a second value threshold. In this case, the information queue of the target object is optimized to obtain an optimized queue. The specific implementation process of S103 may include: S1035-S1036, as follows: S1035. Shorten the length of the coarse queue in the information queue of the target object to obtain a simplified queue.

[0102] When the exposure value of an object within a target group is equal to or lower than the second value threshold, it indicates that these objects are unlikely to convert and are considered low-conversion objects. Generally, the total computing power for information promotion is limited. Allocating the same computing power to low-conversion objects as to high-conversion objects would waste some computing power, thus reducing the likelihood of high-conversion objects converting. Therefore, in this embodiment, the information promotion device deletes some promotional information from the coarse queue of the information queue corresponding to the target object, thereby shortening the length of the coarse queue and obtaining a streamlined queue. This reduces the computing power allocated to low-conversion objects, allowing more computing power to be concentrated on high-conversion objects. Thus, the number of promotional information in the streamlined queue is less than the number of promotional information in the coarse queue.

[0103] S1036. The promotion information in the simplified queue is sorted to obtain the optimized queue.

[0104] The information promotion equipment sends the promotion information in the streamlined queue into the fine ranking model. The finely ranked sequence obtained by the fine ranking model is the optimized queue.

[0105] For example, Figure 6 This is a schematic diagram illustrating the optimization of a queue by shortening the length of the coarse-sorted queue, as provided in an embodiment of this application. Queue 6-1 is the coarse-sorted queue in the information queue. Originally, the length of queue 6-1 was 7. The information promotion device deletes promotional information 6-11 and 6-12 from queue 6-1, resulting in a simplified queue 6-2. Then, the simplified queue 6-2 is fine-sorted, and promotional information is further filtered. The resulting fine-sorted queue is the optimized queue 6-3. From Figure 5 As can be seen, when the information promotion equipment performs fine sorting on the simplified queue 6-2, the computing power required will inevitably be less than that required for queue 6-1.

[0106] In this embodiment, the information promotion device shortens the length of the coarse queue to obtain a streamlined queue for objects whose exposure value is lower than or equal to the second value threshold. Then, it performs fine sorting on the streamlined queue to obtain an optimized queue. Thus, for these low-conversion objects, less computing power can be used to promote information to them, avoiding the waste of computing power and improving the optimization of the information promotion process for low-conversion objects.

[0107] See Figure 7 , Figure 7 This is an optional flowchart illustrating the information promotion method provided in the embodiments of this application. Figure 2 In some embodiments of this application, the target group includes objects with a click-through rate lower than a click-through rate threshold. In this case, the information queue of the target objects is optimized to obtain an optimized queue. The specific implementation process of S103 may include: S1037-S10310, as follows: S1037. Obtain historical search data and historical conversion data of the target object.

[0108] S1038. Using historical search data and historical conversion data, explore the interests of the target object and obtain interest tags.

[0109] Click-through rate (CTR) indicates whether an audience is interested in the promotional information. When the audience includes individuals with CTRs below the threshold, it suggests that these individuals may not be interested in the promoted information and are considered low-interest targets. In this case, the promotional device will retrieve the target audience's historical search data and historical conversion data from other platforms. It will then use this historical search and conversion data to determine what types of products or information the target audience is interested in, thus obtaining their interest tags.

[0110] It should be noted that both historical search data and historical conversion data are actively generated by the target audience and can indicate their interests. Historical search data may include the keywords searched by the target audience, the number of searches, and the product or information type corresponding to the search keywords. Historical conversion data may include the type and name of the product for which the target audience made a conversion, and may also include the value of the product for which the target audience made a conversion; however, this application does not impose any limitations on these aspects.

[0111] It is understood that historical search data and historical conversion data can be the search and conversion data of the target object on the information promotion platform during a historical period, or it can be the search and conversion data of the target object on other platforms, such as search platforms and shopping platforms. This application does not limit these.

[0112] S1039. Select interest-based promotional information from the platform's information pool based on interest tags.

[0113] S10310. Optimize the information queue by promoting information based on interests to obtain an optimized queue.

[0114] After obtaining the target audience's interest tags, the information promotion device can use these tags to select promotional information corresponding to the interest tags from the platform's information pool. This promotional information is called interest-based promotional information. Next, the device will use the selected interest-based promotional information to optimize the target audience's information queue, resulting in an optimized queue of promotional information that better matches the target audience's interests.

[0115] In this embodiment, the information promotion device can first explore the target object's interest tags from its historical search data and historical conversion data. Then, based on these interest tags, it selects interest-based promotional information from the platform's information pool and optimizes the information queue using this selected information to obtain an optimized queue. This increases the probability of pushing promotional information that is of interest to low-interested individuals to the target object, thereby optimizing the information promotion process for low-interested individuals and improving the targeting of information promotion.

[0116] In some embodiments of this application, the information queue is optimized using interest-based promotion information to obtain an optimized queue. The specific implementation process of S10310 may include: S10310a or S10310b, as follows: S10310a. Replace the promotional information in the information queue with interest-based promotional information to obtain an optimized queue.

[0117] Information promotion equipment can use interest-based promotional information to replace any promotional information in the information queue to obtain an optimized queue, or it can calculate the similarity between interest-based promotional information and each promotional information in the information queue, and use interest-based promotional information to replace the promotional information with the lowest similarity to obtain an optimized queue.

[0118] It is understandable that information promotion devices can use interest-based promotional information to replace the promotional information in the coarse-ranked queue of the information queue, and then perform fine-ranking on the replaced coarse-ranked queue to obtain an optimized queue. Alternatively, they can directly replace the promotional information in the fine-ranked queue of the information queue to obtain an optimized queue.

[0119] S10310b: Add interest promotion information to the information queue to obtain the optimized queue.

[0120] The information promotion device can directly insert interest promotion information at the end of the information queue, or insert interest promotion information at any position in the information queue; this application does not limit this.

[0121] Understandably, the information promotion device can add interest-based promotional information to the coarse-ranked queue of the information queue, and then perform fine-ranking on the coarse-ranked queue with the added interest-based promotional information to obtain the optimized queue, or directly add interest-based promotional information to the fine-ranked queue of the information queue to obtain the optimized queue.

[0122] In this embodiment, the information promotion device can optimize the information queue by either replacing the promotional information in the information queue with interest-based promotional information or adding interest-based promotional information to the information queue. This results in an optimized queue, which facilitates the promotion of information of interest to low-interest individuals, thereby optimizing the information promotion process for low-interest individuals.

[0123] In some embodiments of this application, the target group includes: when the click-through rate is lower than the click-through rate threshold, the information queue of the target object is optimized to obtain an optimized queue. The specific implementation process of S103 may also include: S201-S203, as follows: S201. Determine the current industry category to which the promotional information in the target object's information queue belongs.

[0124] Information promotion equipment identifies the industry category of each promotional message in the target audience's information queue. This can be achieved by analyzing industry tags within the promotional message or by using artificial intelligence to analyze the content of the promotional message.

[0125] S202. Select other promotional information from the platform's information pool that is different from the current industry category.

[0126] S203. Add other promotional information to the information queue to obtain the optimization queue.

[0127] Because some industries offer high value during information promotion, the same type of promotional information might be repeatedly returned during the information return phase. This could prevent the return of promotional information that the target audience is truly interested in, or cause them to become annoyed by the returned information. To address this, the information promotion device selects promotional information from the platform's information pool that falls outside the current industry category. These selected promotional information are called "alternative promotional information." The device then adds this alternative promotional information to the end of the information queue, or to any other position, creating an optimized queue.

[0128] It is understandable that the information promotion device can add other promotional information to the coarse queue of the information queue, and then perform fine ranking on the coarse queue after adding the information to obtain the optimized queue, or directly add other promotional information to the fine ranking queue of the information queue. This application does not limit this.

[0129] For example, Figure 8 This is a schematic diagram illustrating how to obtain an optimized queue using other promotional information, as provided in an embodiment of this application. See also... Figure 8 In information queue 8-1, multiple promotional messages are currently categorized under the gaming industry (8-2). Since gaming industry (8-2) generally has higher value in promotional contexts, these messages are easily returned to the target audience. To address this, the promotional device selects promotional messages from other categories (e.g., automotive industry (8-4) and sports industry (8-5)) from the platform's information pool (8-3) and adds them to the end of information queue 8-1. This optimizes the queue, allowing promotional messages from industries other than gaming (8-2) to also have a chance to be pushed to the target audience.

[0130] In this embodiment, the information promotion device can first determine the current industry category of the promotional information in the information queue, and then select promotional information from other industry categories different from the current industry category from the platform information pool, adding the other promotional information to the information queue. This allows promotional information from other industries to also have the opportunity to be pushed to the target audience, increasing the click-through rate of the promotional information and thus achieving targeted optimization of promotional information for low-interest audiences.

[0131] In some embodiments of this application, the target group includes: objects whose exposure count is lower than or equal to a second exposure threshold, and objects whose exposure value is lower than or equal to a second value threshold. In this case, the information promotion method provided in the embodiments of this application may further include: S105, as follows: S105. Reduce the frequency of information promotion to the target audience.

[0132] If an object's exposure count is less than or equal to the second exposure threshold, and its exposure value is less than or equal to the second value threshold, it indicates that the object rarely views promotional information and rarely converts in response to promotional information; therefore, it is a low-value object. Information promotion devices need to protect the experience of low-value objects, which requires reducing the frequency of information promotion targeting these objects to minimize promotional activity and protect their user experience.

[0133] In this embodiment of the application, when the information promotion device clearly identifies the target object as a low-value object, it will also reduce the frequency of information promotion to the target object to protect the user experience and optimize the information promotion process for low-value objects.

[0134] It should be noted that in some embodiments, the target group can be the intersection of the different target groups provided in the above embodiments. Therefore, the target group may be a more precise group, such as objects whose exposure count exceeds a first exposure threshold and whose exposure value exceeds a first value threshold, or objects whose exposure value is equal to or lower than a second value threshold and whose click-through rate is lower than a click-through rate threshold. In this case, the information promotion device can select one or more optimization strategies from the optimization strategies provided in the above embodiments based on the types of objects included in the target group, forming a composite optimization strategy. Then, the composite optimization strategy is used to optimize the information queue of the target group; that is, for a more precise group, multiple different optimization strategies are used simultaneously to optimize the information queue.

[0135] In some embodiments of this application, historical operation data includes the number of exposures, exposure value, and click-through rate corresponding to the target object. In this case, determining the target object's group based on the historical operation data, i.e., the specific implementation process of S102, may include any one or more of S1021-S1023, as follows: S1021. The number of exposures of the target object is compared with the first exposure threshold and the second exposure threshold respectively to determine the object group to which the target object belongs.

[0136] It should be noted that, in the embodiments of this application, the number of exposures represents the number of promotional information viewed by the target object on average per day.

[0137] S1022. The exposure value of the target object is compared with the first value threshold and the second value threshold respectively to determine the object group to which the target object belongs.

[0138] It is understood that, in the embodiments of this application, exposure value represents the value of thousands of exposures of promotional information viewed by the target object.

[0139] S1023. Compare the click-through rate of the target object with the click-through rate threshold to determine the target object group to which the target object belongs.

[0140] In this embodiment of the application, the click-through rate (CTR) represents the click-through rate of promotional information viewed by the target object, which is calculated from the number of clicks and exposures of the promotional information.

[0141] In some embodiments, the information promotion device can determine the target group of the target object by using only any of the processing steps in S1021-S1023, thereby making the scope of the determined target group wider, so that the corresponding information queue can be optimized by only one optimization strategy to obtain the optimization strategy.

[0142] In other embodiments, the information promotion device can also perform S1021, S1022 and S1023 simultaneously for the target object, that is, for the target object, the target group to which it belongs is determined by the number of exposures, exposure value and click rate, so that different optimization strategies can be used to optimize the corresponding information queue at the same time to obtain an optimized queue.

[0143] In this embodiment, the information promotion device can divide the target object into groups based on the number of exposures, exposure value, and click-through rate of the target object in historical operation data, so as to obtain the target object group to which the target object belongs, and to optimize the information queue by using the optimization strategy corresponding to the target group.

[0144] In some embodiments of this application, before determining the object group to which the target object belongs based on historical operation data, i.e., before S101, the method may further include: S106-S108, as follows: S106. Analyze the average daily exposure of objects on the information platform to obtain the first exposure threshold and the second exposure threshold.

[0145] The information promotion equipment obtains the average daily exposure of the objects on the information platform, and then analyzes the average daily exposure of these objects to obtain information such as the mean, median, and mode of the average daily exposure of each object. Then, the mean of the average daily exposure is used as the first exposure threshold, and the median is used as the second exposure threshold.

[0146] It should be noted that the average daily exposure represents the average number of promotional messages viewed daily by the target audience of the information platform. The target audience of the information platform can be all registered users on the platform, or users within the same age group as the target audience; this application does not impose any limitations on this.

[0147] In some embodiments of this application, the information promotion device may also directly determine the first preset exposure amount as the first exposure threshold and the second preset exposure amount as the second exposure threshold. Both the first and second preset exposure amounts can be set according to actual conditions, such as being set to 1 or 2, etc., and this application does not limit this setting.

[0148] S107. Analyze the exposure value of promotional information in the platform's information pool to obtain the first value threshold and the second value threshold.

[0149] In some embodiments, the information promotion device can directly determine the average exposure value of all promotional information in the platform information pool as the first value threshold and the median as the second value threshold. In other embodiments, the information promotion device can also sort the exposure values ​​of promotional information in the platform information pool to select the exposure value of a specific position to obtain the first value threshold and the second value threshold.

[0150] S108. Analyze the click-through rate of promotional information in the platform's information pool to obtain the click-through rate threshold.

[0151] Similarly, the information promotion equipment will also analyze the click-through rate (CTR) of promotional information in the platform's information pool, using the average CTR of all promotional information or the CTR of a specific location as the average CTR. It should be noted that, since CTR data is relatively sparse, in some embodiments, it is necessary to analyze the CTR of promotional information that has reached a certain number of exposures to obtain a more reasonable CTR threshold.

[0152] In this embodiment of the application, the information promotion device will use the daily average exposure of the object on the information platform, the exposure value of the promotion information, and the click-through rate to determine the first exposure threshold, the second exposure threshold, the first value threshold, the second value threshold, and the click-through rate threshold, so as to use these thresholds to stratify the target object in the future.

[0153] In some embodiments of this application, the exposure value of promotional information in the platform information pool is analyzed to obtain a first value threshold and a second value threshold. The specific implementation process of S107 may include: S1071-S1072, as follows: S1071. The exposure value of the promotional information in the platform's information pool is arranged from largest to smallest to obtain the exposure value sequence.

[0154] S1072. The exposure value of the first position in the exposure value sequence is determined as the first value threshold, and the exposure value of the second position in the exposure value sequence is determined as the second value threshold.

[0155] The information promotion equipment sorts the exposure value of each promotional message in the platform's information pool according to their relative values, thereby obtaining an exposure value sequence. Then, it selects the first exposure value from the exposure value sequence as the first value threshold, and then selects the second exposure value from the exposure value sequence as the second value threshold.

[0156] It is understood that the first position can be preset, such as the 10th, 15th, etc., or it can be the 1 / 2nd, 1 / 3rd, etc., of the entire exposure value sequence. This application does not limit this. Similarly, the second position can be preset, or it can be the 3rd, 4th, 5th, etc., of the entire exposure sequence. This application does not limit this.

[0157] In this embodiment of the application, the information promotion device can sort the exposure values ​​of each promotional information by size to obtain an exposure value sequence. Then, based on the exposure value of a specific position in the exposure value sequence, a first value threshold and a second value threshold are determined to facilitate subsequent stratification of the target object.

[0158] In some embodiments of this application, the click-through rate of promotional information in the platform information pool is analyzed to obtain a click-through rate threshold. The specific implementation process of S108 may include: S1081-S1083, as follows: S1081. From the promotional information in the platform's information pool, filter out the promotional information that has been displayed more than the preset number of times to be analyzed.

[0159] The information promotion device first reads the display count of each promotional message in the platform's information pool. The display count indicates the exposure of each promotional message. Then, the device compares the display count of each promotional message with a preset count to filter out promotional messages from the platform's information pool that have been displayed more than the preset count. These promotional messages have been displayed a sufficient number of times, and their click-through rates are reliable.

[0160] It is understandable that the preset number of times can be set according to the actual situation, such as 20 or 60, etc., and this application does not limit it here.

[0161] S1082. Arrange the click-through rates of the promotional information to be analyzed in order of magnitude to obtain the click-through rate sequence.

[0162] S1083. The click rate of the preset position in the click rate sequence is determined as the click rate threshold.

[0163] After receiving the promotional information to be analyzed, the information promotion equipment sorts the click-through rates (CTRs) of the information by size, such as sorting them from largest to smallest or smallest to largest, thus obtaining a CTR sequence. Next, the equipment finds a preset CTR position within this sequence, such as the second-highest CTR or the CTR at the halfway point of the entire sequence, and uses this as a CTR threshold.

[0164] In this embodiment, the information promotion device first filters out promotional information with a confidence in click-through rate from the promotional information in the platform information pool, sorts the promotional information to be analyzed by click-through rate, and determines the click-through rate of the preset position in the sorted click-through rate sequence as the click-through rate threshold, so as to facilitate subsequent segmentation of the target objects.

[0165] The following will describe an exemplary application of the embodiments of this application in a real-world application scenario.

[0166] This application embodiment is implemented in a scenario where a server (information promotion device) recommends advertisements (promotional information) to users (target objects). This application embodiment segments users and then applies user segmentation tags to the ranking model and strategy model of the advertising system.

[0167] First, the server needs to determine the segmentation metrics. The server segments users using impressions (impression data), CPM (value per mille), and CTR (click-through rate). Here, impressions represent user activity, CPM represents the value per thousand impressions, and CTR represents the user's interest in the recommended ads.

[0168] When tiering servers, the following principles should be followed: the CTR metric should be reliable; the percentage of page views and data consumption for each tier should not be too small; and the tiering standards for each site should be as consistent as possible.

[0169] In the specific segmentation process, the server mainly segments users based on their exposure, CPM, and CTR (historical action data) over the past 30 days. The methods for determining the thresholds for each metric during segmentation are as follows: 1. Regarding exposure, it is mainly determined based on physical meaning. For example, an average of 1 post per day is the threshold for low exposure. Those with fewer than 1 post per day are considered low-exposure individuals (the number of exposures is lower than or equal to the second exposure threshold). An average of 2 posts per day is the threshold for high exposure. Those with more than 2 posts per day are considered high-exposure individuals (the number of exposures is higher than the first exposure threshold). Those with more than 1 post per day but less than 2 posts per day are considered intermediate.

[0170] 2. For CPM, the 1 / 2 quantile (i.e., the first rank in the entire CPM ranking) is used as the threshold for high CPM (first value threshold), and the 3 / 4 quantile (i.e., the second rank in the entire CPM ranking) is used as the threshold for low CPM (second value threshold). This is because: See [link to relevant documentation] Figure 9The provided user distribution chart for CPM segments (vertical axis: user percentage 9-1, ranging from 0.00% to 20.00%, and cumulative user percentage 9-2, ranging from 0% to 120%; horizontal axis: average CPM 9-3, ranging from 0 to 100) shows that users of a certain site are mainly concentrated below an average CPM of 15 yuan (left-skewed). Therefore, it is not appropriate to use the average as the standard for judging high or low CPM. It is more suitable to use 1 / 2 and 3 / 4 as the standards for high and low, with those between the 1 / 2 and 3 / 4 quantiles considered as the intermediate state. In this case, those with CPM below the 3 / 4 quantile are low-value users, those with CPM above the 1 / 2 quantile are high-value users, and those with CPM between the 1 / 2 and 3 / 4 quantiles are the intermediate state, i.e., normal users.

[0171] 3. Regarding CTR, due to the sparse click data, CTR is unreliable under low exposure. Therefore, CTR analysis needs to be performed under certain exposure conditions. The server uniformly limits the exposure (number of impressions) to greater than or equal to 60 (preset number of times) as the standard for a reliable CTR. Using the order of reliable CTR from highest to lowest (the click-through rate of the promotional information to be analyzed), the 1 / 2 quantile is used as the threshold for high CTR (first exposure threshold), and the 3 / 4 quantile is used as the threshold for low CTR (second exposure threshold) to segment users. That is, people with CPM below the 3 / 4 quantile are considered low-click users, people with CPM above the 1 / 2 quantile are considered high-click users, and people with CPM above the 3 / 4 quantile but below the 1 / 2 quantile are considered intermediate. Through this segmentation, the standard for low click-through rate is almost 1%, which is consistent with the understanding of information feed click-through rate.

[0172] The server uses the thresholds determined in the above manner to segment users, resulting in several user groups as shown in Table 1.

[0173] Table 1

[0174] In this context, "normal users" refers to users whose CPM, click-through rate (CPM), and CTR are all in the middle range. Based on the above results, the server can determine different optimization directions (optimization strategies corresponding to the target group), as shown in Table 2: Table 2

[0175] The strategies will be introduced below.

[0176] Direct Access to Top-Ranked Ads: Analysis revealed that for the same user, several requests within a short time interval showed high similarity in the ranking queue. Some ads at the very end of the queue consistently remained at the bottom, receiving no exposure. Therefore, in this strategy, the server, targeting users with high exposure (those whose exposure exceeds the first exposure threshold), relinquishes these ads at the end of the ranking queue (promotional information at the end of the ranking queue in the information queue) and moves them to other ads (replacing the promotional information).

[0177] Increased Computational Resources: The advertising system mainly consists of four stages: recall, coarse ranking, fine ranking, and ad return. During recall, ads are matched to users, typically filtering out 20,000 to 50,000 ads. To reduce time consumption, a coarse ranking model is used to reduce the number of ads to 200-300, which are then fed into a fine ranking model for selection. From the selected ads, the one with the highest preset value per thousand impressions is chosen for exposure. Currently, all requests sent to the fine ranking are of the same length, resulting in equal computational resources for high-value users (those with exposure value exceeding a first value threshold) and low-value users. To address this, the server extends the queue length from coarse to fine ranking for high-value users (creating an expanded queue). This allows for more ads to be used for fine ranking of high-value users, helping to filter out ads that are more likely to lead to conversions among high-value users.

[0178] Reduce computing resources: For low-value groups (objects whose exposure value is lower than or equal to the second value threshold), the server shortens the queue length from coarse ranking to fine ranking (resulting in a simplified queue) to reduce computing resources for low-value groups.

[0179] Interest Exploration (Cocoon Audience): Analysis revealed that high-value, low-click audiences (this operation can be performed on low-click users regardless of their high-value status, i.e., those with click-through rates below the threshold) have a high proportion of ads from the gaming industry (current industry category), approximately 80%. These ads are not necessarily of genuine interest to this audience; they are simply pushed to users because of their high value per thousand impressions. Therefore, appropriately adding ads from other industries (other promotional information from other industry categories) to this audience can appropriately improve the click-through rate.

[0180] Interest Exploration (Silent Audience): This segment consists of highly active users who ignore ads. While they experience a high variety of ad exposures and high user activity, they are not likely to click on ads. Therefore, it can be assumed that the recommended ads for this group are not what they prefer. In this case, we can leverage search recommendation data (historical search data) and incorporate cross-site data (historical conversion data from other platforms) to conduct interest exploration (obtaining interest tags to generate interest-based promotional information).

[0181] Reduce exposure: Reduce the frequency of ad recommendations to users (reduce the frequency of information promotion to the target audience). However, periodic exposure should be considered to determine whether the user has become a high-potential audience.

[0182] Add search recommendation data: This is the data recommended when a user searches, and it is usually the data that the user is interested in.

[0183] It should be noted that since each user group is determined by a combination of CPM, exposure, and CTR, meaning that each user group is the intersection of the thresholds that meet these metrics, the optimization direction for each user group is also a set of optimization strategies corresponding to CPM, exposure, and CTR.

[0184] By using the above methods, the queue is optimized according to different optimization strategies for different types of users, so as to recommend advertisements in different ways for different types of users, thereby improving the targeting of information promotion.

[0185] The following description continues to illustrate the exemplary structure of the information promotion device 555 provided in the embodiments of this application as a software module. In some embodiments, such as... Figure 2 As shown, the software modules stored in the information dissemination device 555 in the memory 550 may include: Data acquisition module 5551 is used to acquire historical operation data of the target object; The group determination module 5552 is used to determine the target group to which the target object belongs based on the historical operation data, wherein the target group includes objects that have the same operation mode for the promotional information; The queue optimization module 5553 is used to optimize the information queue of the target object based on the optimization strategy corresponding to the target group to obtain an optimized queue, wherein the optimized queue includes candidate promotion information selected for the target object; The information push module 5554 is used to filter out target promotion information from the optimization queue and push the target promotion information to the target object.

[0186] In some embodiments of this application, the target group includes: objects whose exposure count exceeds a first exposure threshold; the queue optimization module 5553 is further configured to select replacement promotional information from various promotional information in the platform information pool; and use the replacement promotional information to replace the promotional information at the tail of the fine-ranked queue in the information queue of the target object to obtain the optimized queue.

[0187] In some embodiments of this application, the target group includes: objects whose exposure value exceeds a first value threshold; the queue optimization module 5553 is further configured to extend the length of the coarse-order queue in the information queue of the target object to obtain an expanded queue; the number of promotional information in the expanded queue is greater than the number of promotional information in the coarse-order queue; and the promotional information in the expanded queue is fine-ordered to obtain the optimized queue.

[0188] In some embodiments of this application, the target group includes: objects whose exposure value is equal to or lower than a second value threshold; the queue optimization module 5553 is further configured to shorten the length of the coarse queue in the information queue of the target object to obtain a simplified queue; the number of promotional information in the simplified queue is less than the number of promotional information in the coarse queue; and the promotional information in the simplified queue is finely sorted to obtain the optimized queue.

[0189] In some embodiments of this application, the target group includes: objects with a click-through rate lower than a click-through rate threshold; the queue optimization module 5553 is further configured to acquire historical search data and historical conversion data of the target objects; use the historical search data and the historical conversion data to explore the interests of the target objects and obtain interest tags; select interest promotion information from the platform information pool based on the interest tags; and optimize the information queue using the interest promotion information to obtain the optimized queue.

[0190] In some embodiments of this application, the queue optimization module 5553 is further configured to replace the promotional information in the information queue with the interest promotional information to obtain the optimized queue; or, to add the interest promotional information to the information queue to obtain the optimized queue.

[0191] In some embodiments of this application, the queue optimization module 5553 is further configured to determine the current industry category to which the promotional information in the information queue of the target object belongs; select other promotional information from the platform information pool that is different from the current industry category; and add the other promotional information to the information queue to obtain the optimized queue.

[0192] In some embodiments of this application, the object group includes: objects whose exposure count is less than or equal to a second exposure threshold, and objects whose exposure value is less than or equal to a second value threshold; The information promotion device 555 further includes: a frequency adjustment module 5555; The frequency adjustment module 5555 is used to reduce the frequency of information promotion to the target object.

[0193] In some embodiments of this application, the historical operation data includes the number of exposures, exposure value, and click-through rate corresponding to the target object; the group determination module 5552 is further configured to compare the number of exposures of the target object with a first exposure threshold and a second exposure threshold respectively to determine the object group to which the target object belongs; or, compare the exposure value of the target object with a first value threshold and a second value threshold respectively to determine the object group to which the target object belongs; or, compare the click-through rate of the target object with a click-through rate threshold to determine the object group to which the target object belongs.

[0194] In some embodiments of this application, the information promotion device 555 further includes: a threshold determination module 5556; the threshold determination module 5556 is used to analyze the average daily exposure of objects on the information platform to obtain a first exposure threshold and a second exposure threshold; the average daily exposure represents the number of promotional information viewed by objects on the information platform on an average day; analyze the exposure value of promotional information in the platform information pool to obtain a first value threshold and a second value threshold; and analyze the click-through rate of promotional information in the platform information pool to obtain the click-through rate threshold.

[0195] In some embodiments of this application, the threshold determination module 5556 is further configured to arrange the exposure value of the promotional information in the platform information pool from largest to smallest to obtain an exposure value sequence; determine the first exposure value in the exposure value sequence as the first value threshold, and determine the second exposure value in the exposure value sequence as the second value threshold.

[0196] In some embodiments of this application, the threshold determination module 5556 is further configured to: filter out promotional information to be analyzed from the promotional information in the platform information pool whose display count is greater than a preset number; arrange the click-through rates of the promotional information to be analyzed in order of magnitude to obtain a click-through rate sequence; and determine the click-through rate of the preset position in the click-through rate sequence as the click-through rate threshold.

[0197] This application provides a computer program product or computer program that includes 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 information dissemination method described in this application.

[0198] This application provides a computer-readable storage medium storing executable instructions, including executable information dissemination instructions. When these executable instructions are executed by a processor, the processor will execute the information dissemination method provided in this application. For example, ... Figure 3 The method shown.

[0199] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or it may be a variety of devices including one or any combination of the above-mentioned memories.

[0200] In some embodiments, executable information promotion instructions may take the form of programs, software, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

[0201] As an example, executable information promotion instructions may, but do not necessarily, correspond to files in the file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborative files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0202] As an example, executable information promotion instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0203] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. An information promotion method, characterized in that, include: Acquire historical operation data of the target object, including exposure value; The exposure value of promotional information in the platform's information pool is arranged from largest to smallest to obtain an exposure value sequence; the first exposure value in the exposure value sequence is determined as the first value threshold, and the second exposure value in the exposure value sequence is determined as the second value threshold. Based on the historical operation data, the target object is identified to which it belongs, wherein the target object group includes objects that have the same operation pattern in response to promotional information; Based on the optimization strategy corresponding to the target group, the information queue of the target object is optimized to obtain an optimized queue. The optimized queue includes candidate promotional information selected for the target object. When the target group is an object whose exposure value exceeds the first value threshold, promotional information is selected from the platform information pool and added to the coarse-ranking queue of the information queue to obtain an expanded queue. The promotional information in the expanded queue is then input into the fine-ranking model for fine-ranking to obtain the optimized queue. The number of promotional information in the expanded queue is greater than the number of promotional information in the coarse-ranking queue. When the target group is an object whose exposure value is equal to or lower than the second value threshold, promotional information is selected from the coarse-ranked queue of the information queue and deleted to obtain a simplified queue. The promotional information in the simplified queue is then input into the fine-ranking model for fine-ranking to obtain the optimized queue. The number of promotional information in the simplified queue is less than the number of promotional information in the coarse-ranked queue. Target promotion information is selected from the optimization queue and pushed to the target object.

2. The method according to claim 1, characterized in that, The target group includes: objects whose exposure count exceeds a first exposure threshold; the optimization of the information queue of the target objects to obtain an optimized queue includes: Select replacement promotional information from the various promotional information in the platform's information pool; Using the replacement promotion information, the promotion information at the tail of the fine-ranking queue in the information queue of the target object is replaced to obtain the optimized queue.

3. The method according to claim 1, characterized in that, The target group includes: objects with a click-through rate lower than the click-through rate threshold; the optimization of the information queue of the target objects to obtain an optimized queue includes: Obtain the historical search data and historical conversion data of the target object; Using the historical search data and the historical conversion data, interest exploration is performed on the target object to obtain interest tags; Based on the interest tags, interest-based promotional information is selected from the platform's information pool; The information queue is optimized using the interest-based promotion information to obtain the optimized queue.

4. The method according to claim 3, characterized in that, The step of optimizing the information queue using the interest-based promotion information to obtain the optimized queue includes: The optimized queue is obtained by replacing the promotional information in the information queue with the interest-based promotional information; or... The interest promotion information is added to the information queue to obtain the optimization queue.

5. The method according to claim 3, characterized in that, The optimization of the information queue of the target object to obtain an optimized queue includes: Determine the current industry category to which the promotional information in the information queue of the target object belongs; From the platform's information pool, select other promotional information from other industry categories that are different from the current industry category; The other promotional information is added to the information queue to obtain the optimization queue.

6. The method according to claim 1, characterized in that, The target group includes: objects whose exposure count is less than or equal to a second exposure threshold, and objects whose exposure value is less than or equal to a second value threshold; the method further includes: Reduce the frequency of information promotion targeting the aforementioned target group.

7. The method according to any one of claims 1 to 6, characterized in that, The historical operation data includes the number of exposures, exposure value, and click-through rate corresponding to the target object; determining the target object's group based on the historical operation data includes: The number of exposures to the target object is compared with a first exposure threshold and a second exposure threshold to determine the object group to which the target object belongs; or, The exposure value of the target object is compared with the first value threshold and the second value threshold respectively to determine the object group to which the target object belongs; or, The click-through rate of the target object is compared with a click-through rate threshold to determine the object group to which the target object belongs.

8. The method according to claim 7, characterized in that, Before determining the target object's group based on the historical operation data, the method further includes: The average daily exposure of objects on the information platform is analyzed to obtain the first exposure threshold and the second exposure threshold; the average daily exposure represents the amount of promotional information viewed by the objects on the information platform on average per day; The click-through rate of promotional information in the platform's information pool is analyzed to obtain the click-through rate threshold.

9. The method according to claim 8, characterized in that, The step of analyzing the click-through rate of promotional information in the platform's information pool to obtain the click-through rate threshold includes: From the promotional information in the platform's information pool, select promotional information that has been displayed more than a preset number of times to be analyzed; The click-through rates of the promotional information to be analyzed are arranged in order of magnitude to obtain a click-through rate sequence; The click rate at a preset position in the click rate sequence is determined as the click rate threshold.

10. An information promotion device, characterized in that, include: The data acquisition module is used to acquire historical operation data of the target object, including exposure value; The threshold determination module is used to sort the exposure value of the promotional information in the platform information pool from largest to smallest to obtain an exposure value sequence; the first exposure value in the exposure value sequence is determined as the first value threshold, and the second exposure value in the exposure value sequence is determined as the second value threshold. The group determination module is used to determine the target group to which the target object belongs based on the historical operation data, wherein the target group includes objects that have the same operation mode for the promotional information; The queue optimization module is used to optimize the information queue of the target object based on the optimization strategy corresponding to the target group, to obtain an optimized queue. The optimized queue includes candidate promotional information selected for the target object. When the target group is an object whose exposure value exceeds the first value threshold, the queue optimization module is further used to select promotional information from the platform information pool and add it to the coarse-ranked queue of the information queue to obtain an expanded queue. The promotional information in the expanded queue is then input into the fine-ranking model for fine-ranking to obtain the optimized queue. The number of promotional information in the expanded queue is greater than the number of promotional information in the coarse-ranked queue. When the target group is an object whose exposure value is equal to or lower than the second value threshold, the queue optimization module is further used to select promotional information from the coarse-ranked queue of the information queue for deletion, to obtain a simplified queue. The promotional information in the simplified queue is then input into the fine-ranking model for fine-ranking to obtain the optimized queue. The number of promotional information in the simplified queue is less than the number of promotional information in the coarse-ranked queue. The information push module is used to filter out target promotion information from the optimization queue and push the target promotion information to the target object.

11. The apparatus according to claim 10, characterized in that, The target group includes: objects whose exposure count exceeds a first exposure threshold; The queue optimization module is also used to select replacement promotional information from various promotional information in the platform information pool; Using the replacement promotion information, the promotion information at the tail of the fine-ranking queue in the information queue of the target object is replaced to obtain the optimized queue.

12. The apparatus according to claim 10, characterized in that, The target group includes: objects whose click-through rate is lower than the click-through rate threshold; The queue optimization module is also used to acquire historical search data and historical conversion data of the target object; Using the historical search data and the historical conversion data, interest exploration is performed on the target object to obtain interest tags; Based on the aforementioned interest tags, interest-based promotional information is selected from the platform's information pool; The information queue is optimized using the interest-based promotion information to obtain the optimized queue.

13. An information promotion device, characterized in that, include: Memory, used to store executable information and instructions; A processor, when executing executable information extension instructions stored in the memory, implements the method according to any one of claims 1 to 9.

14. A computer-readable storage medium, characterized in that, The device stores executable information promotion instructions, which, when executed by a processor, implement the method described in any one of claims 1 to 9.

15. A computer program product comprising computer-executable information dissemination instructions or a computer program, characterized in that, When the computer executes the information promotion instructions or the computer program is executed by the processor, it implements the method according to any one of claims 1 to 9.

Images