Method, device, computer equipment and storage medium for pushing content

Abstract

The present application provides a method, apparatus, computer device, and storage medium for pushing content, which can be applied to the fields of cloud computing, artificial intelligence, or blockchain, and is used to solve the problem of low accuracy in pushing content. The method comprises: determining a target tag combination associated with a target account based on the account characteristics of a target account, wherein the target tag combination is used to characterize the account type of the target account; determining a target push factor corresponding to the target tag combination based on a mapping relationship between a pre-stored tag combination and a push factor, wherein the push factor is obtained based on the quality assessment value of the corresponding tag combination and resource consumption reference information; and pushing target content to the target account based on the obtained target push factor.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, apparatus, computer device, and storage medium for pushing content. Background Art

[0002] With the continuous development of technology, devices can perform targeted content push tasks. For example, devices can push different content to different accounts.

[0003] Before the device performs a targeted content push task, it needs to predict the target account's expected revenue per thousand impressions (eCPM) for different content. The device pushes the target content with the highest eCPM value to the target account based on the predicted eCPM value. However, when pushing content based solely on the eCPM value, it may happen that the target account has high eCPM values ​​for multiple contents, but the target account is more likely to consume resources for only one of the multiple contents. If one of the multiple contents is randomly pushed to the target account, it will affect the content provider's return on investment (ROI). It can be seen that in the process of pushing content to the target account, the accuracy of the pushed content is low. Summary of the Invention

[0004] The embodiments of the present application provide a method, apparatus, computer device, and storage medium for pushing content, which are used to solve the problem of low accuracy of pushed content.

[0005] In a first aspect, a method for pushing content is provided, comprising:

[0006] Determining a target tag combination associated with the target account based on the account characteristics of the target account, wherein the target tag combination is used to characterize the account type of the target account;

[0007] Determine a target push factor corresponding to the target tag combination based on a pre-stored mapping relationship between tag combinations and push factors, wherein the push factor is obtained based on a quality assessment value of the corresponding tag combination and resource consumption reference information, wherein the quality assessment value represents the probability of resource consumption by each sample account associated with the corresponding tag combination, and the resource consumption reference information represents fused information of historical resource consumption records corresponding to each sample account associated with the corresponding tag combination;

[0008] Based on the obtained target push factor, the target content is pushed to the target account.

[0009] In a second aspect, a method for transposing pushed content is provided, including:

[0010] A first processing module is configured to determine a target tag combination associated with the target account based on the account characteristics of the target account, wherein the target tag combination is used to characterize the account type of the target account;

[0011] The first processing module is further configured to determine a target push factor corresponding to the target tag combination based on a pre-stored mapping relationship between tag combinations and push factors, wherein the push factor is obtained based on a quality evaluation value of the corresponding tag combination and resource consumption reference information, wherein the quality evaluation value represents the probability of resource consumption of each sample account associated with the corresponding tag combination, and the resource consumption reference information represents fused information of historical resource consumption records corresponding to each sample account associated with the corresponding tag combination;

[0012] The second processing module is configured to push the target content to the target account based on the obtained target push factor.

[0013] Optionally, the first processing module is specifically configured to:

[0014] Determine a conversion target group index corresponding to the label combination based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of conversion sample accounts associated with the label combination, and the total number of conversion sample accounts associated with each pre-stored label combination, wherein the conversion sample account is a sample account that has consumed resources among the sample accounts associated with the label combination;

[0015] Determine a non-conversion target group index corresponding to the label combination based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of non-conversion sample accounts associated with the label combination, and the total number of non-conversion sample accounts associated with each pre-stored label combination, wherein the non-conversion sample account is a sample account that has not consumed resources among the sample accounts associated with the label combination;

[0016] Based on the obtained conversion target group index and non-conversion target group index, a quality assessment value associated with the tag combination is determined.

[0017] Optionally, the processing module is specifically configured to:

[0018] Fusing the quality evaluation value of the label combination and the resource consumption reference information to determine a reinforcement coefficient corresponding to the label combination;

[0019] If the conversion target group index corresponding to the tag combination is greater than a first preset threshold, mapping the reinforcement coefficient corresponding to the tag combination to a first specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped reinforcement coefficient;

[0020] If the non-conversion target group index corresponding to the tag combination is greater than a second preset threshold, mapping the reinforcement coefficient corresponding to the tag combination to a second specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped reinforcement coefficient;

[0021] Any value in the first specified interval is greater than any value in the second specified interval.

[0022] Optionally, the first processing module is specifically configured to:

[0023] Based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, a plurality of preliminary label combinations are obtained, and preliminary sample account sets corresponding to each of the plurality of preliminary label combinations are obtained;

[0024] Based on the obtained quality assessment values ​​of each of the multiple preliminary label combinations, screening out the preliminary label combinations whose quality assessment values ​​meet a preset first selection condition as the selected label combinations, and using the preliminary sample account sets corresponding to the obtained respective selected label combinations as the selected sample account sets of the corresponding label combinations;

[0025] For each selected tag combination obtained, the following operations are respectively performed: based on the historical resource consumption records corresponding to each selected sample account in the selected sample account set associated with one tag combination in each selected tag combination, the historical resource consumption records corresponding to the selected sample account that has consumed resources in the selected sample account set are integrated to obtain resource consumption reference information for the one tag combination;

[0026] Based on the obtained quality evaluation values ​​of the respective selected tag combinations and the resource consumption reference information, push factors corresponding to the respective selected tag combinations are determined to obtain the mapping relationship.

[0027] Optionally, the first processing module is further configured to:

[0028] Before obtaining a plurality of preliminary label combinations based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, and obtaining the preliminary sample account sets corresponding to each of the plurality of preliminary label combinations, obtaining each sample account in response to a selection operation triggered on the client;

[0029] Based on the historical resource consumption records corresponding to each of the sample accounts, sample accounts that have consumed resources are screened out from the sample accounts as conversion sample accounts, and the other sample accounts are screened out as non-conversion sample accounts;

[0030] The sample account set is determined based on the obtained conversion sample accounts and non-conversion sample accounts.

[0031] Optionally, the first processing module is specifically configured to:

[0032] Based on the account characteristics of each sample account in the sample account set, determining the preset tag associated with each sample account from the pre-stored preset tags;

[0033] Inputting the sample accounts and the preset tags associated with the sample accounts into the trained tag selection model, screening out the preset tags that meet the second selection condition, and determining a plurality of preliminary tag combinations based on the screened preset tags;

[0034] Based on the sample accounts associated with each of the multiple preliminary label combinations, preliminary sample account sets associated with corresponding preliminary label combinations are determined respectively, and preliminary sample account sets associated with each of the multiple preliminary label combinations are obtained.

[0035] Optionally, the processing module is specifically configured to:

[0036] Based on the conversion target group index corresponding to each of the preliminary tag combinations, screening out the preliminary tag combinations whose conversion target group index is greater than a first preset threshold as the selected tag combination;

[0037] Based on the non-conversion target group index corresponding to each of the preliminary tag combinations, screening out preliminary tag combinations whose non-conversion target group index is greater than a second preset threshold as the selected tag combination;

[0038] Based on the obtained respective selected label combinations, a plurality of label combinations are obtained.

[0039] According to a third aspect, a computer device is provided, comprising:

[0040] a memory for storing program instructions;

[0041] The processor is configured to call the program instructions stored in the memory and execute the method described in the first aspect according to the obtained program instructions.

[0042] In a fourth aspect, a storage medium is provided, wherein the storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the method described in the first aspect.

[0043] In an embodiment of the present application, a quality assessment value is used to characterize the probability of each sample account associated with the corresponding label combination consuming resources, so that an evaluation can be made of the ability of each sample account associated with the corresponding label combination to consume resources. Resource consumption reference information is used to characterize the fusion information of the historical resource consumption records of each sample account associated with the corresponding label combination, so that the total amount of resources consumed by each sample account associated with the corresponding label combination can be obtained. In the mapping relationship, the push factor is obtained based on the quality assessment value of the corresponding label combination and the resource consumption reference information, so that the push factor can characterize the possibility of each sample account associated with the corresponding label combination consuming resources. Therefore, when pushing target content to the target account based on the target push factor, the content with the greatest possibility of consuming resources can be selected as the target content, thereby increasing the possibility of the target account consuming resources for the target content, and then improving the income-output ratio of the content provider, so as to achieve the purpose of improving the accuracy of the pushed content. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 An application scenario of the method for pushing content provided in an embodiment of the present application;

[0045] Figure 2 A flow chart of a method for pushing content provided in an embodiment of the present application Figure 1 ;

[0046] Figure 3 A schematic diagram of the principle of the method for pushing content provided in the embodiment of the present application Figure 1 ;

[0047] Figure 4a A flow chart of a method for pushing content provided in an embodiment of the present application Figure 2 ;

[0048] Figure 4b A schematic diagram of the principle of the method for pushing content provided in the embodiment of the present application Figure 2 ;

[0049] Figure 4c A schematic diagram of the principle of the method for pushing content provided in the embodiment of the present application Figure 3 ;

[0050] Figure 4d A fourth schematic diagram of a method for pushing content provided in an embodiment of the present application;

[0051] Figure 4e A fifth schematic diagram of a method for pushing content provided in an embodiment of the present application;

[0052] Figure 5aAn interface diagram of the method for pushing content provided in the embodiment of the present application Figure 2 ;

[0053] Figure 5b An interface diagram of the method for pushing content provided in the embodiment of the present application Figure 2 ;

[0054] Figure 6 An interactive diagram of the method for pushing content provided in an embodiment of the present application;

[0055] Figure 7 Schematic diagram of the structure of the device for pushing content provided in the embodiment of the present application Figure 1 ;

[0056] Figure 8 Schematic diagram of the structure of the device for pushing content provided in the embodiment of the present application Figure 2 . DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0058] Some of the terms used in the embodiments of the present application are explained below to facilitate understanding by those skilled in the art.

[0059] (1) Expected revenue per thousand impressions (eCPM):

[0060] eCPM refers to the revenue earned per thousand impressions. The unit of impression can be a webpage, a content unit, or a single piece of content. Based on eCPM, the device can display the content with the highest eCPM to the account. Factors influencing eCPM include estimated click-through rate, estimated conversion rate, and the content provider's target bid.

[0061] (2) Return on Investment (ROI):

[0062] "Revenue" refers to the total static investment in a project, while "output" refers to the total value added over the project's entire operational lifespan. ROI is the ratio of the project's total investment to the total value added of output over its operational lifespan. A smaller ROI indicates a better economic outcome.

[0063] The embodiments of this application involve cloud technology, artificial intelligence (AI), and blockchain. They are designed based on cloud computing and cloud storage in cloud technology, and based on machine learning (ML) in artificial intelligence.

[0064] Cloud technology refers to a hosting technology that unifies hardware, software, and network resources within a wide or local area network (WAN) to enable data computing, storage, processing, and sharing. Cloud technology is a general term for network, information technology, integration technology, management platform technology, and application technology, all based on the cloud computing business model. It can form a resource pool for on-demand, flexible, and convenient use. Cloud computing technology will become a crucial support. Backend services for technical network systems, such as those for video sites, image sites, and more portals, require significant computing and storage resources. With the rapid development and application of the internet industry, every item will likely have its own unique identifier, requiring transmission to backend systems for logical processing. Data of varying levels will be processed separately, and data from all industries will require robust system support, which can only be achieved through cloud computing.

[0065] Cloud computing is a computing model that distributes computing tasks across a resource pool consisting of a large number of computers, enabling various application systems to access computing power, storage space, and information services as needed. The network that provides these resources is called the "cloud." To users, these resources appear infinitely scalable and can be accessed at any time, used on demand, expanded at any time, and paid for on a pay-per-use basis.

[0066] As a provider of cloud computing infrastructure, a cloud computing resource pool (referred to as a cloud platform, generally referred to as an Infrastructure as a Service (IaaS) platform) is established. Various types of virtual resources are deployed in the resource pool for external customers to choose and use. The cloud computing resource pool mainly includes: computing devices (virtualized machines, including operating systems), storage devices, and network devices.

[0067] Based on logical functional divisions, the Platform as a Service (PaaS) layer can be deployed on top of the IaaS layer, and the Software as a Service (SaaS) layer can be deployed on top of the PaaS layer. SaaS can also be deployed directly on top of IaaS. PaaS is a platform for software execution, such as databases and web containers. SaaS is a variety of business software, such as web portals and text messaging apps. Generally speaking, SaaS and PaaS are layers above IaaS.

[0068] Cloud storage is a new concept that has been extended and developed from the concept of cloud computing. A distributed cloud storage system (hereinafter referred to as a storage system) refers to a storage system that uses cluster applications, grid technology, and distributed storage file systems to bring together a large number of different types of storage devices (storage devices are also called storage nodes) in the network through application software or application interfaces to work together and provide external data storage and business access functions.

[0069] Currently, storage systems utilize a method for creating logical volumes. When creating a logical volume, physical storage space is allocated for each logical volume. This physical storage space may consist of disks on a specific storage device or several storage devices. When a client stores data on a logical volume, it stores the data on a file system. The file system divides the data into multiple parts, each of which is an object. An object contains not only the data but also additional information such as the data identifier (ID entity). The file system writes each object to the physical storage space of the logical volume and records the storage location information for each object. Therefore, when a client requests access to data, the file system can provide access based on the storage location information for each object.

[0070] The storage system allocates physical storage space to logical volumes by pre-dividing the physical storage space into stripes based on the estimated capacity of the objects to be stored in the logical volume (this estimate often has a large margin relative to the actual capacity of the objects to be stored) and the Redundant Array of Independent Disks (RAID) groupings. A logical volume can be understood as a stripe, thereby allocating physical storage space to the logical volume.

[0071] Artificial intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive field of computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also studies the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making. AI technologies primarily encompass computer vision, natural language processing, machine learning, and deep learning.

[0072] With the research and advancement of artificial intelligence technology, artificial intelligence has been studied and applied in many fields, such as common smart homes, smart recommendation systems, virtual assistants, smart speakers, smart marketing, smart translation, autonomous driving, robots, smart medical care, etc. It is believed that with the development of technology, artificial intelligence will be applied in more fields and play an increasingly important role.

[0073] Machine learning is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. It specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.

[0074] Blockchain is a new application model for computer technologies, including distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a series of data blocks linked using cryptographic methods. Each block contains information about a batch of online transactions, used to verify the validity of this information (to prevent counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, the platform product and service layer, and the application service layer.

[0075] The underlying blockchain platform can include processing modules such as user management, basic services, smart contracts, and operation monitoring. Among them, the user management module is responsible for the identity information management of all blockchain participants, including maintaining public and private key generation (account management), key management, and maintaining the corresponding relationship between the user's real identity and the blockchain address (authority management), etc., and under authorization, it supervises and audits the transactions of certain real identities and provides risk control rule configuration (risk control audit); the basic service module is deployed on all blockchain node devices to verify the validity of business requests, and records the valid requests to the storage after consensus is reached. For a new business request, the basic service first adapts the interface to parse and authenticate the request (interface adaptation), and then encrypts the business information through the consensus algorithm (consensus management). The smart contract module is responsible for the registration, issuance, triggering and execution of contracts. Developers can define the contract logic in a programming language and publish it to the blockchain (contract registration). According to the logic of the contract terms, the contract logic is triggered by calling keys or other events to trigger execution. The contract logic is completed, and the contract upgrade and cancellation functions are also provided. The operation monitoring module is mainly responsible for the deployment, configuration modification, contract setting, cloud adaptation and real-time status visualization output of the product during the product release process, such as alarms, network status monitoring, and node device health monitoring.

[0076] The platform's product service layer provides the basic capabilities and implementation framework for typical applications. Developers can build on these basic capabilities, overlay business features, and complete the blockchain implementation of business logic. The application service layer provides application services based on blockchain solutions for business participants to use.

[0077] The following is a brief introduction to the application fields of the method for pushing content provided in the embodiments of the present application.

[0078] With the continuous development of technology, more and more content providers are pushing content to various accounts through devices. Since the number of content that a device can push to a target account at the same time or at the same node is limited, the device can push the content with the highest eCPM value to the target account based on the expected revenue per thousand impressions (eCPM) predicted by each content provider for the target account and content.

[0079] When pushing content, the device can perform targeted content push tasks. For example, the device can push different content to different accounts. When pushing content based solely on eCPM values, it's possible that the target account's eCPM values ​​for multiple pieces of content are all high, but the target account is likely to consume resources more for only one piece of content. If a random piece of content is pushed to the target account, this will affect the content provider's return on investment (ROI). Therefore, when pushing content to the target account, the accuracy of the pushed content is low.

[0080] In order to solve the problem of low accuracy of pushed content, the present application proposes a method for pushing content. The method determines the target tag combination associated with the target account based on the account characteristics of the target account. After obtaining the target tag combination associated with the target account, the target push factor corresponding to the target tag combination is determined based on the mapping relationship between the pre-stored tag combination and the push factor. After obtaining the target push factor corresponding to the target tag combination, the target content is pushed to the target account based on the target push factor. Among them, in the mapping relationship, the push factor is obtained based on the quality evaluation value of the corresponding tag combination and the resource consumption reference information, the quality evaluation value is used to characterize the probability of each sample account associated with the corresponding tag combination consuming resources, and the resource consumption reference information is used to characterize the fusion information of the historical resource consumption records corresponding to each sample account associated with the corresponding tag combination.

[0081] In an embodiment of the present application, a quality assessment value is used to characterize the probability of each sample account associated with the corresponding label combination consuming resources, so that an evaluation can be made of the ability of each sample account associated with the corresponding label combination to consume resources. Resource consumption reference information is used to characterize the fusion information of the historical resource consumption records of each sample account associated with the corresponding label combination, so that the total amount of resources consumed by each sample account associated with the corresponding label combination can be obtained. In the mapping relationship, the push factor is obtained based on the quality assessment value of the corresponding label combination and the resource consumption reference information, so that the push factor can characterize the possibility of each sample account associated with the corresponding label combination consuming resources. Therefore, when pushing target content to the target account based on the target push factor, the content with the greatest possibility of consuming resources can be selected as the target content, thereby increasing the possibility of the target account consuming resources for the target content, and then improving the income-output ratio of the content provider, so as to achieve the purpose of improving the accuracy of the pushed content.

[0082] The following describes the application scenarios of the method for pushing content provided by this application.

[0083] Please refer to Figure 1, which is an application scenario of the method for pushing content provided in this application. The application scenario includes a client 101 and a server 102. The client 101 includes a first client 1011 and a second client 1012. The first client 1011 and the server 102 can communicate with each other, and the second client 1012 and the server 102 can communicate with each other. The communication method can be to use wired communication technology, such as communicating by connecting a network cable or a serial cable; or to use wireless communication technology, such as communicating by Bluetooth or wireless fidelity (WIFI) and other technologies, without specific limitation.

[0084] Client 101 generally refers to a device that can provide content to server 102, such as a terminal device, a third-party application accessible by the terminal device, or a webpage accessible by the terminal device. Examples of terminal devices include mobile phones, tablet computers, and personal computers. Server 102 generally refers to a device that can push content to an account, such as a terminal device or a server. Examples of servers include cloud servers and local servers. Both client 101 and server 102 can utilize cloud computing to reduce the use of local computing resources; similarly, cloud storage can be used to reduce the use of local storage resources.

[0085] As an embodiment, the client 101 and the server 102 can be the same device, and there is no specific limitation. In the embodiment of the present application, the client 101 and the server 102 are different devices as an example for introduction.

[0086] The following is based on Figure 1 , the method for pushing content provided in the embodiment of the present application is specifically introduced, taking the server as the server 102 as an example.

[0087] Please refer to Figure 2 , which is a flow chart of the method for pushing content provided in an embodiment of the present application.

[0088] S201: The server determines a target tag combination associated with the target account based on the account characteristics of the target account.

[0089] There are various situations in which the server determines the target tag combination associated with the target account based on the account characteristics of the target account. For example, the server determines the target tag combination associated with the target account when responding to the login operation of the target account; for another example, the server determines the target tag combination associated with the target account when responding to the target account accessing the target node; for another example, the server determines the target tag combination associated with the target account when the access duration of the target account is greater than the preset duration, etc.

[0090] The target node may be a target website, a target interface, a target link, a target Internet Protocol (IP) address, etc. The preset duration may be user-defined or calculated by the server based on recent account access duration, etc., without limitation.

[0091] The server can read the target account's account information and extract the target account's account characteristics. For example, the server can read the target account's age, gender, associated game accounts, or historical resource consumption records to determine the target account's account characteristics. Based on the target account's account characteristics, the server determines the preset tags associated with the target account from among the pre-stored preset tags. For example, if the target account's age is 26, then one of the preset tags associated with the target account is determined to be {age: [25, 30]}.

[0092] The server determines the target tag combination associated with the target account from among the pre-stored tag combinations based on the preset tags associated with the target account. When each tag in the tag combination matches the preset tag associated with the target account, the tag combination can be determined to be the target tag combination associated with the target account. For example, the preset tags associated with the target account include {age: [25, 30]} and {gender: male}, and the pre-stored tag combinations include a first tag combination of {age: [25, 30]} and a second tag combination of {age: [20, 25]} and {gender: male}, then the target tag combination associated with the target account is determined to be the first tag combination. When, among the pre-stored tag combinations, multiple tag combinations match the preset tags associated with the target account, the tag combination including the largest number of tags can be determined to be the target tag combination associated with the target account. For example, the preset tags associated with the target account include {age: [25, 30]} and {gender: male}, and the pre-stored tag combinations include the first tag combination of {age: [25, 30]} and the second tag combination of {age: [25, 30]} and {gender: male}, then the target tag combination associated with the target account is determined to be the second tag combination.

[0093] S202: The server determines a target push factor corresponding to a target tag combination based on a pre-stored mapping relationship between tag combinations and push factors.

[0094] After obtaining the target tag combination associated with the target account, the server can determine the target push factor corresponding to the target tag combination from the pre-stored mapping relationship between the tag combination and the push factor. The mapping relationship between the tag combination and the push factor can be a mapping relationship determined by the server for different content, or a mapping relationship determined by the server for different types of content, or a mapping relationship determined by the server for content provided by different clients 101, etc., or a mapping relationship received by the server from other devices, without specific limitation. In the embodiment of the present application, the mapping relationship determined by the server for content provided by different clients 101 is used as an example for introduction.

[0095] The push factor in the mapping relationship is obtained based on the quality evaluation value of the corresponding label combination and the resource consumption reference information. The quality evaluation value of the corresponding label combination is obtained using the following method (detailed introduction to the method of obtaining the mapping relationship later):

[0096] Based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of conversion sample accounts associated with the label combination, and the total number of conversion sample accounts associated with each pre-stored label combination, the conversion target group index corresponding to the label combination is determined, wherein the conversion sample account is the sample account that has consumed resources among the sample accounts associated with the label combination. Based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of non-conversion sample accounts associated with the label combination, and the total number of non-conversion sample accounts associated with each pre-stored label combination, the non-conversion target group index corresponding to the label combination is determined, wherein the non-conversion sample account is the sample account that has not consumed resources among the sample accounts associated with the label combination. Based on the obtained conversion target group index and non-conversion target group index, a quality assessment value associated with the label combination is determined.

[0097] The push factor of the corresponding tag combination in the mapping relationship is obtained by the following method (the method of obtaining the mapping relationship will be described in detail later):

[0098] The quality assessment value of the label combination and the resource consumption reference information are fused to determine the reinforcement coefficient corresponding to the label combination. If the conversion target group index corresponding to the label combination is greater than the first preset threshold, the reinforcement coefficient corresponding to the label combination is mapped to the first specified interval, and the push factor corresponding to the label combination is obtained based on the mapped reinforcement coefficient. If the non-conversion target group index corresponding to the label combination is greater than the second preset threshold, the reinforcement coefficient corresponding to the label combination is mapped to the second specified interval, and the push factor corresponding to the label combination is obtained based on the mapped reinforcement coefficient. Among them, any numerical value in the first specified interval is greater than any numerical value in the second specified interval. Among them, the first preset threshold can be pre-set according to the usage scenario, or it can be calculated based on the conversion target group index corresponding to each label combination, and there is no specific restriction. The second preset threshold can be pre-set according to the usage scenario, or it can be calculated based on the conversion target group index corresponding to each label combination, and there is no specific restriction. The first preset threshold and the second preset threshold can be the same or different.

[0099] The following describes a method for obtaining a mapping relationship between a label combination and a push factor.

[0100] Please refer to Figure 3 , which is a schematic diagram illustrating the principle of a method for obtaining a mapping relationship between a label combination and a push factor. The server can determine multiple preliminary label combinations based on the account characteristics of each sample account in the sample account set and pre-stored preset labels, and obtain preliminary sample account sets corresponding to each of the multiple preliminary label combinations. Based on the historical resource consumption records corresponding to the preliminary sample account sets associated with each of the multiple preliminary label combinations, the server can determine the percentage of preliminary sample accounts that have consumed resources in each of the preliminary sample account sets, and based on the obtained percentage information corresponding to each of the preliminary sample account sets, determine the quality assessment value of each preliminary label combination. The server can obtain each quality assessment value and, from each preliminary label combination, select preliminary label combinations whose quality assessment values ​​meet a preset first selection condition as selected label combinations, and use the preliminary sample account sets corresponding to each selected label combination as the selected sample account sets for the corresponding label combination. The server can determine the push factor corresponding to each selected label combination based on the historical resource consumption records corresponding to each selected sample account in the selected sample account set associated with each selected label combination. The server may obtain a mapping relationship based on the push factors corresponding to the obtained respective selected tag combinations.

[0101] Please refer to Figure 4a , which is a flowchart of a method for obtaining a mapping relationship between a label combination and a push factor.

[0102] S401: The server determines each preliminary tag combination.

[0103] Please refer to Figure 4b , which is a schematic diagram illustrating the principle of obtaining the mapping relationship between label combinations and push factors. Based on the obtained conversion sample accounts and non-conversion sample accounts, the server can determine a sample account set. It then inputs each sample account in the sample account set, along with the preset labels associated with each sample account, into the trained label selection model to filter out the preset labels that meet the second selection criteria, thereby obtaining a preliminary label combination. This is described in detail below.

[0104] Before obtaining multiple preliminary tag combinations based on the account characteristics of each sample account in the sample account set and the pre-stored preset tags, the server may first obtain the sample account set through client 101, receive the sample account set from another device, or read a pre-stored sample account set. There are multiple ways for the server to obtain the sample account set, two of which are described below as examples.

[0105] Method 1:

[0106] In response to the selection operation triggered on the client 101 , the server obtains each conversion sample account and each non-conversion sample account, and determines a sample account set based on the obtained each conversion sample account and each non-conversion sample account.

[0107] The content provider can perform a selection operation through the first client 1011 or the second client 1012. In response to the selection operation for the first client 1011 or the second client 1012, the server obtains each conversion sample account and each non-conversion sample account. The server determines a sample account set based on the obtained conversion sample accounts and each non-conversion sample account.

[0108] For example, the content provider can search on the first client 1011 based on historical resource consumption records to retrieve sample accounts that have consumed resources, or retrieve sample accounts that have consumed resources at the target node. The content provider can select all sample accounts that have consumed resources through the first client 1011, and can select all sample accounts that have consumed resources at the target node. In response to the content provider's selection operation on the first client 1011, the server obtains each converted sample account and treats other unselected sample accounts as non-converted sample accounts. It can also select sample accounts that have not consumed resources at the target node as non-converted sample accounts to obtain each non-converted sample account. The server can determine a sample account set based on each converted sample account and each non-converted sample account obtained.

[0109] Method 2:

[0110] In response to the selection operation triggered on the client 101 , the server obtains each sample account, and based on the obtained sample accounts, determines each conversion sample account and each non-conversion sample account to obtain a sample account set.

[0111] The content provider can select each sample account through the first client 1011 or the second client 1012. The selection operation can be a select-all operation or a search operation. For example, the sample accounts with female gender in the account information can be selected; another example is the sample accounts with Beijing as the location in the account information; another example is the sample accounts with female gender and Beijing as the location in the account information, etc. This makes the sample accounts obtained more targeted and the mapping relationship based on the sample accounts more accurate.

[0112] After obtaining each sample account in response to a selection operation on the first client 1011 or the second client 1012, the server can filter out sample accounts that have consumed resources from each sample account based on the historical resource consumption records corresponding to each sample account as conversion sample accounts, or filter out sample accounts that have consumed resources at the target node as conversion sample accounts; at the same time, the server can treat other sample accounts as non-conversion sample accounts, or treat sample accounts that have not consumed resources at the target node as non-conversion sample accounts, thereby obtaining each conversion sample account and each non-conversion sample account. The server can determine a sample account set based on the obtained each conversion sample account and each non-conversion sample account.

[0113] Method 3:

[0114] The server determines each conversion sample account and each non-conversion sample account based on the pre-stored historical resource consumption records corresponding to each sample account, and obtains a sample account set.

[0115] The server can read the historical resource consumption records of all sample accounts and filter out the sample accounts that have consumed resources as conversion sample accounts. It can also filter out the sample accounts that have consumed resources at the target node as conversion sample accounts. At the same time, it can treat other sample accounts as non-conversion sample accounts, or it can treat the sample accounts that have not consumed resources at the target node as non-conversion sample accounts. Based on the historical resource consumption records of all sample accounts, the server can filter out the sample accounts that have consumed resources within a preset time period as conversion sample accounts, and treat the sample accounts that have never consumed resources as non-conversion sample accounts, etc. After obtaining each conversion sample account and each non-conversion sample account, the server determines a sample account set based on each conversion sample account and each non-conversion sample account.

[0116] As an example, each sample account in the sample account set can be represented by an account ID. For example, an account ID uniquely represents each sample account in the sample account set. Conversion sample accounts and non-conversion sample accounts can be represented by attribute identifiers. For example, a conversion attribute of 0 for a sample account indicates that the sample account is a non-conversion sample account, while a conversion attribute of 1 for a sample account indicates that the sample account is a conversion sample account.

[0117] As an embodiment, when performing a search, the search conditions may also include multiple conditions, such as the time of consuming resources or the amount of consumed resources.

[0118] After obtaining the sample account set, the server can determine the preset tags associated with each sample account from among the pre-stored preset tags based on the account characteristics of each sample account in the sample account set. The process by which the server determines the preset tags associated with each sample account is similar to the process by which the server determines the preset tags associated with the target account in S201 and is not further described here.

[0119] After obtaining the preset tags associated with each sample account, the server inputs each sample tag and the preset tags associated with each sample account into the trained tag selection model. The trained tag selection model can filter out preset tags that meet the second selection criteria, or filter out second-order combinations of preset tags that meet the second selection criteria, or filter out third-order combinations of preset tags that meet the second selection criteria, etc. Filtering out preset tags that meet the second selection criteria indicates that the selected preset tags are of high importance, i.e., the sample accounts associated with the selected preset tags are more likely to consume resources for content provided by the content provider.

[0120] The server may determine a preliminary tag combination based on the preset tags screened out by the trained tag selection model. For example, the server may treat each preset tag screened out as a preliminary tag combination, treat every two preset tags as a preliminary tag combination, treat every three preset tags as a preliminary tag combination, and so on, to obtain multiple preliminary tag combinations. The server may determine a preliminary tag combination based on the second-order combination of preset tags screened out by the trained tag selection model. For example, the server may treat each preset tag in the second-order combination of preset tags screened out as a preliminary tag combination, treat the second-order combination of each preset tag as a preliminary tag combination, and so on, to obtain multiple preliminary tag combinations. The server may determine a preliminary tag combination based on the third-order combination of preset tags screened out by the trained tag selection model. For example, the server may treat each preset tag in the third-order combination of preset tags screened out as a preliminary tag combination, treat every two preset tags in the third-order combination of preset tags as a preliminary tag combination, treat the third-order combination of each preset tag as a preliminary tag combination, and so on, to obtain multiple preliminary tag combinations. The method for determining the preliminary tag combination is not specifically limited.

[0121] As an embodiment, the content provider can specify the screening range of the trained tag selection model in each pre-stored preset tag through the first client 1011 or the second client 1012. The content provider can specify a preset tag, and the trained tag selection model will use the specified preset tag as the primary tag combination. The trained tag selection model can also filter out the second-order combination of preset tags including the specified preset tag, and the trained tag selection model can also filter out the third-order combination of preset tags including the specified preset tag, etc. Please refer to Figure 5a For example, if the specified preset label is age, then when the age is specified as [0, 15], the trained label selection model can use [0, 15] as the primary label combination, or the second-order combination of the preset labels including [0, 15] as the primary label combination, etc.

[0122] The content provider can also specify multiple preset tags. The trained tag selection model will then determine the initial tag combinations based on the specified multiple preset tags. The trained tag selection model can also filter out the second-order combinations of preset tags that include the specified multiple preset tags. The trained tag selection model can also filter out the third-order combinations of preset tags that include the specified multiple preset tags. Please refer to Figure 5bFor example, the specified preset labels are age and gender. When the age is specified as [0, 15] and the gender is male, the trained label selection model can use [0, 15] and male as the preliminary label combination, or the third-order combination including the preset labels [0, 15] and male as the preliminary label combination, or the second-order combination including the preset label [0, 15] as the preliminary label combination, and the second-order combination including the preset label male as the preliminary label combination, etc.

[0123] As an embodiment, the trained label selection model is obtained by training the label selection model to be trained based on the various sample accounts obtained by the server and the preset labels associated with each sample account. The label selection model to be trained can calculate the information gain of the combination of each preset label based on the preset labels associated with each obtained sample account and whether each sample account is a converted sample account or a non-converted sample account, and learn the importance of each preset label. Different decision trees are automatically generated according to the importance of each preset label. The depth of the decision tree can be set according to actual usage. Through continuous learning and training of the label selection model to be trained, a trained label selection model can be obtained. The trained label selection model can determine the combination of preset labels with the largest information gain, and the order of the combination is the same as the depth of the decision tree. The label selection model is, for example, a gradient boosting (extreme gradient boosting, XGBoost) model.

[0124] After obtaining each preliminary label combination, the server uses the sample account associated with the preliminary label combination as the preliminary sample account set associated with the preliminary label combination, and obtains the preliminary sample account set associated with each preliminary label combination.

[0125] S402: The server determines the quality evaluation value of each preliminary tag combination.

[0126] After obtaining each preliminary label combination and the preliminary sample account set associated with each preliminary label combination, the server determines the percentage of preliminary sample accounts that have consumed resources in the corresponding preliminary sample account set based on the historical resource consumption records corresponding to the preliminary sample account set associated with each preliminary label combination. The server can also determine the percentage of preliminary sample accounts that have consumed resources in the corresponding preliminary sample account set based on the number of converted sample accounts and the number of non-converted sample accounts in the preliminary sample account set associated with each preliminary label combination. The server can count the values ​​of the conversion attributes of each preliminary sample account to obtain the number of converted sample accounts and the number of non-converted sample accounts, etc.

[0127] The proportion information may include a conversion target group index and a non-conversion target group index. The conversion target group index is used to represent the strength of the conversion sample account in the initial sample account set, while the non-conversion target group index is used to represent the strength of the non-conversion sample account in the initial sample account set.

[0128] Please refer to Figure 4c , the following describes a process of determining the proportion information for one of the multiple preliminary label combinations and determining the quality assessment value based on the proportion information.

[0129] The server may calculate the number of converted sample accounts A included in the preliminary sample account set associated with the one preliminary label combination, the total number T of preliminary sample accounts included in the preliminary sample account sets associated with all preliminary label combinations, and the number A of converted sample accounts included in the preliminary sample account set associated with the one preliminary label combination. + , and the total number of conversion sample accounts included in the primary sample account sets associated with all primary label combinations T + , determine the conversion target group index corresponding to the primary label combination, please refer to formula (1).

[0130]

[0131] The server may calculate the number of preliminary sample accounts A included in the preliminary sample account set associated with the one preliminary label combination, the total number T of preliminary sample accounts included in the preliminary sample account sets associated with all preliminary label combinations, the number A of non-converted sample accounts included in the preliminary sample account set associated with the one preliminary label combination, and the number of non-converted sample accounts A included in the preliminary sample account set associated with the one preliminary label combination. - , and the total number of non-converted sample accounts included in the primary sample account sets associated with all primary label combinations T _ , determine the non-conversion target group index corresponding to the primary label combination, please refer to formula (2).

[0132]

[0133] After the server obtains the conversion target group index corresponding to the preliminary label combination and the non-conversion target group index corresponding to the preliminary label combination, it is equivalent to obtaining the proportion information corresponding to the preliminary label combination. Based on the conversion target group index corresponding to the preliminary label combination and the non-conversion target group index corresponding to the preliminary label combination, the server determines the quality assessment value corresponding to the preliminary label combination. The quality assessment value can be a weighted sum of the conversion target group index and the non-conversion target group index, or it can be the average of the conversion target group index and the non-conversion target group index, etc. The quality assessment value is used to evaluate the probability of each preliminary sample account in the set of preliminary sample accounts associated with the preliminary label combination consuming resources.

[0134] As an embodiment, the proportion information may further include the information gain corresponding to the preliminary label combination. The information gain corresponding to the preliminary label combination is obtained based on the information gain of each preset label combination calculated by the trained label selection model. The information gain corresponding to the preliminary label combination may be the sum of the information gains of all preset labels included in the preliminary label combination, or the weighted sum of the information gains of all preset labels included in the preliminary label combination. The server uses the proportion information corresponding to the preliminary label combination as the quality assessment value of the preliminary label combination.

[0135] S403: The server determines each selected tag combination.

[0136] After obtaining the quality assessment values ​​corresponding to each preliminary label combination, please refer to Figure 4d , the server can filter out the preliminary label combinations that meet the first selection condition as the selected label combination. The first selection condition includes multiple conditions. For example, the server can filter out the preliminary label combinations whose quality evaluation values ​​are greater than the preset quality threshold as the selected label combination. For another example, the server can also filter out the preliminary label combinations whose conversion target group index is greater than the first preset threshold, and the preliminary label combinations whose non-conversion target group index is greater than the second preset threshold, as the selected label combinations, etc. The first preset threshold and the second preset threshold can be referred to the introduction in the previous article and will not be repeated here.

[0137] The server may also sort the various preliminary label combinations in descending order based on the quality evaluation value, and screen out the preliminary label combinations that are ranked before the first preset serial number as the selected label combination. The server may also sort the various preliminary label combinations in descending order based on the conversion target group index, and screen out the preliminary label combinations that are ranked before the second preset serial number, and sort the various preliminary label combinations in descending order based on the non-conversion target group index, and screen out the preliminary label combinations that are ranked before the third preset serial number as the selected label combination. Among them, the first preset serial number can be pre-set according to the usage scenario, or it can be calculated according to the conversion target group index corresponding to each label combination, and there is no specific limitation. The second preset serial number can be pre-set according to the usage scenario, or it can be calculated according to the conversion target group index corresponding to each label combination, and there is no specific limitation. The first preset serial number and the second preset serial number can be the same or different.

[0138] The server uses the preliminary account set associated with the selected label combination as the selected sample account set associated with the corresponding label combination, and obtains the selected sample account set associated with each selected label combination.

[0139] S404: The server determines the push factors corresponding to the respective selected tag combinations and obtains a mapping relationship.

[0140] After obtaining each selected tag combination, the server determines the push factor corresponding to each selected tag combination. The following is an introduction to one of the selected tag combinations. Please refer to Figure 4e .

[0141] The server obtains resource consumption reference information of the tag combination based on the historical resource consumption records corresponding to each selected sample account in the selected sample account set associated with the tag combination and the historical resource consumption records corresponding to the selected sample accounts that have consumed resources in the selected sample account set.

[0142] The method of integrating historical resource consumption records may be to summarize each historical resource consumption record, or to summarize the historical resource consumption records of each selected sample account that has consumed resources. If the historical resource consumption record includes the historical resource consumption amount, then the historical resource consumption amount of each selected sample account that has consumed resources may be summed up. If the historical resource consumption record includes the historical resource consumption interval, then the historical resource consumption interval of each selected sample account that has consumed resources may be intersected, etc., or an average value may be calculated for the historical resource consumption interval of each selected sample account that has consumed resources, and the average values ​​may be summed up, etc. There are no specific restrictions.

[0143] As an example, some content is only pushed to designated nodes. For example, "People You May Know" is only pushed to the Add Friends interface; another example is avatar skins only pushed to the Dress Up interface. Therefore, to further improve the accuracy of pushed content, the server can also integrate historical resource consumption records corresponding to selected sample accounts in the selected sample account set that have consumed resources at the target node to obtain resource consumption reference information for that tag combination. The target node can be a target website, target interface, target industry, or target content display location, etc., without limitation.

[0144] After obtaining the resource consumption reference information for the tag combination, the server determines a push factor corresponding to the tag combination, combining it with the quality assessment value of the tag combination. The server fuses the quality assessment value and the resource consumption reference information for the tag combination to obtain a corresponding enhancement factor. The enhancement factor is used to represent the predicted resource consumption information for each selected sample account in the set of selected sample accounts associated with the tag combination.

[0145] When the resource consumption reference information is a resource consumption reference amount, the server can determine the enhancement coefficient corresponding to the label combination based on the product between the quality evaluation value f of the label combination and the resource consumption reference amount g of the label combination. Please refer to formula (3).

[0146]

[0147] When the resource consumption reference information is a resource consumption reference interval, the server may determine a resource consumption reference average value based on the resource consumption reference interval. The server determines a reinforcement coefficient corresponding to the label combination based on the product of the quality assessment value of the label combination and the resource consumption reference average value of the label combination.

[0148] After obtaining the reinforcement coefficients corresponding to each tag combination, the server may map the reinforcement coefficients to different specified intervals, and aggregate the reinforcement coefficients in the same specified interval into a specified number of categories to obtain corresponding push factors.

[0149] For example, if the conversion target group index corresponding to the label combination is greater than the first preset threshold, then the reinforcement coefficient corresponding to the label combination is mapped to the first specified interval; if the non-conversion target group index corresponding to the label combination is greater than the second preset threshold, then the reinforcement coefficient corresponding to the label combination is mapped to the second specified interval.

[0150] For another example, if the conversion target group index corresponding to the label combination is before the second preset serial number, then the reinforcement coefficient corresponding to the label combination is mapped to the first specified interval; if the non-conversion target group index corresponding to the label combination is before the third preset serial number, then the reinforcement coefficient corresponding to the label combination is mapped to the second specified interval.

[0151] The server aggregates the reinforcement coefficients mapped to the first specified interval into a first specified number of categories to obtain corresponding push factors. The server aggregates the reinforcement coefficients mapped to the second specified interval into a second specified number of categories. The server maps the reinforcement coefficients not mapped to the first specified area and the second specified area to preset values ​​to obtain corresponding push factors.

[0152] For example, the enhancement coefficients mapped to the first specified interval are grouped into three categories, including 1.1, 1.2, and 1.3. The enhancement coefficients mapped to the second specified interval are grouped into three categories, including 0.7, 0.8, and 0.9. The enhancement coefficients not mapped to the first and second specified areas are mapped to 1.

[0153] After obtaining the push factors corresponding to each tag combination, the server establishes a mapping relationship between the tag combination and the push factor. Different content providers can have different mapping relationships, different content can have different mapping relationships, and different types of content can have different mapping relationships, etc., without limitation here.

[0154] S203: The server pushes the target content to the target account based on the obtained target push factor.

[0155] When different content providers push content to a target account through the server, the server can determine the push factors corresponding to the target account based on the corresponding mapping relationship. The server then determines the target content based on each push factor and pushes the target content to the target account.

[0156] The method for the server to determine the target content according to each push factor may be that the server calculates the product between the push factor and the corresponding eCPM, and uses the content provided by the content provider corresponding to the maximum product result as the target content.

[0157] The following is an example of the beneficial effects of the method for pushing content provided in the embodiment of the present application. Please refer to Figure 6 .

[0158] S601 : A first content provider provides first content to a server via a first client 1011 , and a second content provider provides second content to the server via a second client 1012 .

[0159] S602: The server responds to the login operation of the target account on the terminal device.

[0160] S603: Based on the account characteristics of the target account, the server determines the preset tags associated with the target account. The server determines the target account's eCPM for the first content provided by the first content provider, obtaining a first eCPM (e.g., 60). The server also determines the target account's eCPM for the second content provided by the second content provider, obtaining a second eCPM (e.g., 50).

[0161] However, if the target account has recently consumed a large amount of resources for the second content, pushing the content to the target account based on the difference between the first eCPM and the second eCPM will result in inaccurate push.

[0162] S604, the server determines a first target tag combination that matches each preset tag associated with the target account among each tag combination pre-stored for the first content provider, and determines a second target tag combination that matches each preset tag associated with the target account among each tag combination pre-stored for the second content provider.

[0163] S605. The server determines a first target push factor corresponding to the first target tag combination in a mapping relationship pre-stored for the first content provider, for example, 0.8, and determines a second target push factor corresponding to the second target tag combination in a mapping relationship pre-stored for the second content provider, for example, 1.1.

[0164] S606 , the server calculates the product of the first eCPM and the first target push factor to obtain a first result, which is 48, and calculates the product of the second eCPM and the second target push factor to obtain a second result, which is 55.

[0165] S607: The server compares the first result with the second result. If the first result is smaller than the second result, the server pushes the second content to the target account. Therefore, after adjusting the push factor, the content pushed to the target account by the server is more accurate.

[0166] Based on the same inventive concept, the embodiment of the present application provides a device for pushing content, which is equivalent to the server 102 discussed above and can implement the functions corresponding to the aforementioned method for pushing content. Figure 7 , the apparatus includes a first processing module 701 and a second processing module 702, wherein:

[0167] A first processing module 701 is configured to determine a target tag combination associated with the target account based on the account characteristics of the target account, wherein the target tag combination is used to represent the account type of the target account;

[0168] The first processing module 701 is further configured to determine a target push factor corresponding to a target tag combination based on a pre-stored mapping relationship between tag combinations and push factors, wherein the push factor is obtained based on a quality evaluation value of the corresponding tag combination and resource consumption reference information, wherein the quality evaluation value represents the probability of resource consumption by each sample account associated with the corresponding tag combination, and the resource consumption reference information represents fused information of historical resource consumption records corresponding to each sample account associated with the corresponding tag combination.

[0169] The second processing module 702 is configured to push target content to a target account based on the obtained target push factor.

[0170] In a possible embodiment, the first processing module 701 is specifically configured to:

[0171] Determine the conversion target group index corresponding to the label combination based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of conversion sample accounts associated with the label combination, and the total number of conversion sample accounts associated with each pre-stored label combination. The conversion sample account is the sample account that has consumed resources among the sample accounts associated with the label combination.

[0172] Based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of non-converting sample accounts associated with the label combination, and the total number of non-converting sample accounts associated with each pre-stored label combination, the non-converting target group index corresponding to the label combination is determined, where the non-converting sample account is a sample account that has not consumed resources among the sample accounts associated with the label combination;

[0173] Based on the obtained conversion target group index and non-conversion target group index, a quality assessment value of the tag combination association is determined.

[0174] In a possible embodiment, the processing module is specifically configured to:

[0175] The quality evaluation value of the label combination and the resource consumption reference information are integrated to determine the enhancement coefficient corresponding to the label combination;

[0176] If the conversion target group index corresponding to the tag combination is greater than a first preset threshold, mapping the enhancement coefficient corresponding to the tag combination to a first specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped enhancement coefficient;

[0177] If the non-conversion target group index corresponding to the tag combination is greater than a second preset threshold, mapping the reinforcement coefficient corresponding to the tag combination to a second specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped reinforcement coefficient;

[0178] Any value in the first specified interval is greater than any value in the second specified interval.

[0179] In a possible embodiment, the first processing module 701 is specifically configured to:

[0180] Based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, a plurality of preliminary label combinations are obtained, and preliminary sample account sets corresponding to each of the plurality of preliminary label combinations are obtained;

[0181] Based on the quality evaluation values ​​of the multiple preliminary label combinations obtained, the preliminary label combinations whose quality evaluation values ​​meet the preset first selection condition are screened out from the preliminary label combinations as the selected label combinations, and the preliminary sample account sets corresponding to the selected label combinations are respectively used as the selected sample account sets of the corresponding label combinations;

[0182] For each selected label combination obtained, the following operations are performed: based on the historical resource consumption records corresponding to each selected sample account in the selected sample account set associated with one label combination in each selected label combination, the historical resource consumption records corresponding to the selected sample account that has consumed resources in the selected sample account set are integrated to obtain resource consumption reference information for the label combination;

[0183] Based on the obtained quality evaluation values ​​of the respective selected tag combinations and the resource consumption reference information, push factors corresponding to the respective selected tag combinations are determined to obtain a mapping relationship.

[0184] In a possible embodiment, the first processing module 701 is further configured to:

[0185] Before obtaining a plurality of preliminary label combinations based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, and obtaining the preliminary sample account sets corresponding to each of the plurality of preliminary label combinations, in response to a selection operation triggered on the client, obtaining each sample account;

[0186] Based on the historical resource consumption records corresponding to each sample account, sample accounts that have consumed resources are selected as conversion sample accounts, and the other sample accounts are regarded as non-conversion sample accounts;

[0187] A sample account set is determined based on the obtained conversion sample accounts and non-conversion sample accounts.

[0188] In a possible embodiment, the first processing module 701 is specifically configured to:

[0189] Based on the account characteristics of each sample account in the sample account set, determine the preset tag associated with each sample account from the pre-stored preset tags;

[0190] Input each sample account and its associated preset label into the trained label selection model, filter out the preset labels that meet the second selection condition, and determine multiple preliminary label combinations based on the filtered preset labels;

[0191] Based on the sample accounts associated with each of the multiple preliminary label combinations, preliminary sample account sets associated with the corresponding preliminary label combinations are determined respectively, and preliminary sample account sets associated with each of the multiple preliminary label combinations are obtained.

[0192] In a possible embodiment, the processing module is specifically configured to:

[0193] Based on the conversion target group index corresponding to each of the preliminary tag combinations, the preliminary tag combinations having a conversion target group index greater than a first preset threshold are screened out as the selected tag combinations;

[0194] Based on the non-conversion target group index corresponding to each of the preliminary tag combinations, the preliminary tag combinations having a non-conversion target group index greater than a second preset threshold are screened out as the selected tag combinations;

[0195] Based on the obtained respective selected label combinations, a plurality of label combinations are obtained.

[0196] Based on the same inventive concept, an embodiment of the present application provides a computer device, and the computer device 800 is introduced below.

[0197] Please refer to Figure 8 The above-mentioned device for pushing content can run on a computer device 800, and the current version and historical version of the program for pushing content and the application software corresponding to the program for pushing content can be installed on the computer device 800. The computer device 800 includes a display unit 840, a processor 880 and a memory 820, wherein the display unit 840 includes a display panel 841 for displaying a user interactive operation interface, etc.

[0198] In a possible embodiment, the display panel 841 may be configured in the form of a liquid crystal display (LCD) or an organic light-emitting diode (OLED).

[0199] The processor 880 is configured to read a computer program and then execute the method defined by the computer program. For example, the processor 880 reads a program or file for pushing content, thereby running the program for pushing content on the computer device 800 and displaying a corresponding interface on the display unit 840. The processor 880 may include one or more general-purpose processors and may also include one or more DSPs (Digital Signal Processors) to perform related operations to implement the technical solutions provided in the embodiments of the present application.

[0200] The memory 820 generally includes internal memory and external memory. The internal memory can be a random access memory (RAM), a read-only memory (ROM), and a cache (CACHE), etc. The external memory can be a hard disk, an optical disk, a USB disk, a floppy disk or a tape drive, etc. The memory 820 is used to store computer programs and other data. The computer program includes an application corresponding to each client, etc. Other data may include data generated after the operating system or application is run, and the data includes system data (such as configuration parameters of the operating system) and user data. In the embodiment of the present application, program instructions are stored in the memory 820, and the processor 880 executes the program instructions stored in 820 to implement any of the methods for pushing content discussed in the previous figure.

[0201] The display unit 840 is used to receive input digital information, character information, or contact touch operations / contactless gestures, and to generate signal inputs related to user settings and function control of the computer device 800. Specifically, in the embodiment of the present application, the display unit 840 may include a display panel 841. The display panel 841, such as a touch screen, can collect user touch operations on or near it (such as operations performed by the user using a finger, stylus, or any other suitable object or accessory on or on the display panel 841) and drive corresponding connected devices according to a pre-set program.

[0202] In one possible embodiment, the display panel 841 may include a touch detection device and a touch controller. The touch detection device detects the player's touch position and the signal generated by the touch operation, and transmits the signal to the touch controller. The touch controller receives the touch information from the touch detection device, converts it into touch point coordinates, and then sends it to the processor 880. The touch controller can also receive and execute commands from the processor 880.

[0203] The display panel 841 can be implemented using various types such as resistive, capacitive, infrared, and surface acoustic wave. In addition to the display unit 840, the computer device 800 can also include an input unit 830. The input unit 830 can include a graphics input device 831 and other input devices 832. The other input devices can include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control keys, power keys, etc.), a trackball, a mouse, a joystick, etc.

[0204] In addition to the above, the computer device 800 may also include a power supply 890 for powering other modules, an audio circuit 860, a near-field communication module 870, and an RF circuit 810. The computer device 800 may also include one or more sensors 850, such as an accelerometer, a light sensor, a pressure sensor, etc. The audio circuit 860 specifically includes a speaker 861 and a microphone 862. For example, the computer device 800 can use the microphone 862 to collect the user's voice and perform corresponding operations.

[0205] As an embodiment, the number of processors 880 may be one or more, and the processor 880 and the memory 820 may be coupled or relatively independent.

[0206] As an example, Figure 8 The processor 880 in the embodiment can be used to implement the following Figure 7 The functions of the first processing module 701 and the second processing module 702 in FIG.

[0207] As an example, Figure 8 The processor 880 in can be used to implement the corresponding functions of the server 102 discussed above.

[0208] Those skilled in the art will appreciate that all or part of the steps of the above-mentioned method embodiments may be implemented by hardware associated with program instructions, and the aforementioned program may be stored in a computer-readable storage medium. When the program is executed, the program executes the steps of the above-mentioned method embodiments. The aforementioned storage medium includes various media that can store program codes, such as mobile storage devices, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0209] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0210] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for pushing content, characterized in that: include: For multiple content providers, perform the following operations: Based on the account characteristics of the target account, determining a target tag combination associated with the target account from various tag combinations pre-stored for the content provider, wherein the target tag combination is used to characterize the account type of the target account; Determining a target push factor corresponding to the target tag combination based on a mapping relationship between tag combinations and push factors pre-stored for the content provider, wherein the push factor is obtained by fusing a quality assessment value of the corresponding tag combination and resource consumption reference information, the quality assessment value representing the probability of resource consumption by each sample account associated with the corresponding tag combination, and the resource consumption reference information representing fused information of historical resource consumption records corresponding to each sample account associated with the corresponding tag combination; Based on the target push factors respectively obtained for the multiple content providers, target content is selected from the content provided by each of the multiple content providers, and the target content is pushed to the target account.

2. The method according to claim 1, characterized in that The quality assessment value of the label combination is obtained using the following method: Determine a conversion target group index corresponding to the label combination based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of conversion sample accounts associated with the label combination, and the total number of conversion sample accounts associated with each pre-stored label combination, wherein the conversion sample account is a sample account that has consumed resources among the sample accounts associated with the label combination; Determine a non-conversion target group index corresponding to the label combination based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of non-conversion sample accounts associated with the label combination, and the total number of non-conversion sample accounts associated with each pre-stored label combination, wherein the non-conversion sample account is a sample account that has not consumed resources among the sample accounts associated with the label combination; Based on the obtained conversion target group index and non-conversion target group index, a quality assessment value associated with the tag combination is determined.

3. The method according to claim 2, characterized in that The push factor of a tag combination is obtained using the following method: Fusing the quality evaluation value of the label combination and the resource consumption reference information to determine a reinforcement coefficient corresponding to the label combination; If the conversion target group index corresponding to the tag combination is greater than a first preset threshold, mapping the reinforcement coefficient corresponding to the tag combination to a first specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped reinforcement coefficient; If the non-conversion target group index corresponding to the tag combination is greater than a second preset threshold, mapping the reinforcement coefficient corresponding to the tag combination to a second specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped reinforcement coefficient; Any value in the first specified interval is greater than any value in the second specified interval.

4. The method according to any one of claims 1 to 3, characterized in that The mapping relationship is obtained in the following manner: Based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, a plurality of preliminary label combinations are obtained, and preliminary sample account sets corresponding to each of the plurality of preliminary label combinations are obtained; Based on the obtained quality assessment values ​​of each of the multiple preliminary label combinations, screening out the preliminary label combinations whose quality assessment values ​​meet a preset first selection condition as the selected label combinations, and using the preliminary sample account sets corresponding to the obtained respective selected label combinations as the selected sample account sets of the corresponding label combinations; For each selected tag combination obtained, the following operations are respectively performed: based on the historical resource consumption records corresponding to each selected sample account in the selected sample account set associated with one tag combination in each selected tag combination, the historical resource consumption records corresponding to the selected sample account that has consumed resources in the selected sample account set are integrated to obtain resource consumption reference information for the one tag combination; Based on the obtained quality evaluation values ​​of the respective selected tag combinations and the resource consumption reference information, push factors corresponding to the respective selected tag combinations are determined to obtain the mapping relationship.

5. The method according to claim 4, characterized in that Before obtaining a plurality of preliminary label combinations based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, and obtaining preliminary sample account sets corresponding to each of the plurality of preliminary label combinations, the method further includes: In response to a selection operation triggered on the client, obtaining each sample account; Based on the historical resource consumption records corresponding to each of the sample accounts, sample accounts that have consumed resources are screened out from the sample accounts as conversion sample accounts, and the other sample accounts are screened out as non-conversion sample accounts; The sample account set is determined based on the obtained conversion sample accounts and non-conversion sample accounts.

6. The method according to claim 4, characterized in that Based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, a plurality of preliminary label combinations are obtained, and preliminary sample account sets corresponding to each of the plurality of preliminary label combinations are obtained, including: Based on the account characteristics of each sample account in the sample account set, determining the preset tag associated with each sample account from the pre-stored preset tags; Inputting the sample accounts and the preset tags associated with the sample accounts into the trained tag selection model, screening out the preset tags that meet the second selection condition, and determining a plurality of preliminary tag combinations based on the screened preset tags; Based on the sample accounts associated with each of the multiple preliminary label combinations, preliminary sample account sets associated with corresponding preliminary label combinations are determined respectively, and preliminary sample account sets associated with each of the multiple preliminary label combinations are obtained.

7. The method according to claim 4, characterized in that Based on the obtained quality evaluation values ​​of the plurality of preliminary label combinations, screening out, from the respective preliminary label combinations, a preliminary label combination whose quality evaluation value satisfies a preset first selection condition as a selected label combination, including: Based on the conversion target group index corresponding to each of the preliminary tag combinations, screening out the preliminary tag combinations whose conversion target group index is greater than a first preset threshold as the selected tag combination; Based on the non-conversion target group index corresponding to each of the preliminary tag combinations, screening out preliminary tag combinations whose non-conversion target group index is greater than a second preset threshold as the selected tag combination; Based on the obtained respective selected label combinations, a plurality of label combinations are obtained.

8. A device for pushing content, characterized in that: include: The first processing module is used to perform the following operations on multiple content providers: Based on the account characteristics of the target account, determining a target tag combination associated with the target account from various tag combinations pre-stored for the content provider, wherein the target tag combination is used to characterize the account type of the target account; Determining a target push factor corresponding to the target tag combination based on a mapping relationship between tag combinations and push factors pre-stored for the content provider, wherein the push factor is obtained by fusing a quality assessment value of the corresponding tag combination and resource consumption reference information, the quality assessment value representing the probability of resource consumption by each sample account associated with the corresponding tag combination, and the resource consumption reference information representing fused information of historical resource consumption records corresponding to each sample account associated with the corresponding tag combination; The second processing module is configured to select target content from the content provided by each of the multiple content providers based on the target push factors respectively obtained for the multiple content providers, and push the target content to the target account.

9. The device according to claim 8, characterized in that The first processing module is specifically configured to: Determine a conversion target group index corresponding to the label combination based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of conversion sample accounts associated with the label combination, and the total number of conversion sample accounts associated with each pre-stored label combination, wherein the conversion sample account is a sample account that has consumed resources among the sample accounts associated with the label combination; Determine a non-conversion target group index corresponding to the label combination based on the number of sample accounts associated with the label combination, the total number of sample accounts associated with each pre-stored label combination, the number of non-conversion sample accounts associated with the label combination, and the total number of non-conversion sample accounts associated with each pre-stored label combination, wherein the non-conversion sample account is a sample account that has not consumed resources among the sample accounts associated with the label combination; Based on the obtained conversion target group index and non-conversion target group index, a quality assessment value associated with the tag combination is determined.

10. The device according to claim 8, characterized in that The first processing module is specifically configured to: Fusing the quality evaluation value of the label combination and the resource consumption reference information to determine a reinforcement coefficient corresponding to the label combination; If the conversion target group index corresponding to the tag combination is greater than a first preset threshold, mapping the reinforcement coefficient corresponding to the tag combination to a first specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped reinforcement coefficient; If the non-conversion target group index corresponding to the tag combination is greater than a second preset threshold, mapping the reinforcement coefficient corresponding to the tag combination to a second specified interval, and obtaining the push factor corresponding to the tag combination based on the mapped reinforcement coefficient; Any value in the first specified interval is greater than any value in the second specified interval.

11. The device according to claim 8, characterized in that The first processing module is specifically configured to: Based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, a plurality of preliminary label combinations are obtained, and preliminary sample account sets corresponding to each of the plurality of preliminary label combinations are obtained; Based on the obtained quality assessment values ​​of each of the multiple preliminary label combinations, screening out the preliminary label combinations whose quality assessment values ​​meet a preset first selection condition as the selected label combinations, and using the preliminary sample account sets corresponding to the obtained respective selected label combinations as the selected sample account sets of the corresponding label combinations; For each selected tag combination obtained, the following operations are respectively performed: based on the historical resource consumption records corresponding to each selected sample account in the selected sample account set associated with one tag combination in each selected tag combination, the historical resource consumption records corresponding to the selected sample account that has consumed resources in the selected sample account set are integrated to obtain resource consumption reference information for the one tag combination; Based on the obtained quality evaluation values ​​of the respective selected tag combinations and the resource consumption reference information, push factors corresponding to the respective selected tag combinations are determined to obtain the mapping relationship.

12. The device according to claim 11, characterized in that The first processing module is further configured to: Before obtaining a plurality of preliminary label combinations based on the account characteristics of each sample account in the sample account set and each pre-stored preset label, and obtaining the preliminary sample account sets corresponding to each of the plurality of preliminary label combinations, obtaining each sample account in response to a selection operation triggered on the client; Based on the historical resource consumption records corresponding to each of the sample accounts, sample accounts that have consumed resources are screened out from the sample accounts as conversion sample accounts, and the other sample accounts are screened out as non-conversion sample accounts; The sample account set is determined based on the obtained conversion sample accounts and non-conversion sample accounts.

13. The device according to claim 11, characterized in that The first processing module is specifically configured to: Based on the account characteristics of each sample account in the sample account set, determining the preset tag associated with each sample account from the pre-stored preset tags; Inputting the sample accounts and the preset tags associated with the sample accounts into the trained tag selection model, screening out the preset tags that meet the second selection condition, and determining a plurality of preliminary tag combinations based on the screened preset tags; Based on the sample accounts associated with each of the multiple preliminary label combinations, preliminary sample account sets associated with corresponding preliminary label combinations are determined respectively, and preliminary sample account sets associated with each of the multiple preliminary label combinations are obtained.

14. The device according to claim 11, characterized in that The processing module is specifically used for: Based on the conversion target group index corresponding to each of the preliminary tag combinations, screening out the preliminary tag combinations whose conversion target group index is greater than a first preset threshold as the selected tag combination; Based on the non-conversion target group index corresponding to each of the preliminary tag combinations, screening out preliminary tag combinations whose non-conversion target group index is greater than a second preset threshold as the selected tag combination; Based on the obtained respective selected label combinations, a plurality of label combinations are obtained.

15. A computer device, characterized in that: include: a memory for storing program instructions; The processor is configured to call the program instructions stored in the memory, and execute the method according to any one of claims 1 to 7 according to the obtained program instructions.

16. A storage medium, characterized in that The storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

Images