Target account determination method and device, storage medium and electronic equipment

Abstract

The invention discloses a target account determination method and device, a storage medium and electronic equipment. The method comprises the following steps: constructing an account interaction network by utilizing interaction behavior data generated by each object account in an application platform; the account interaction network is divided, a plurality of account communities are obtained, and the association intimacy between object accounts contained in the same account community is larger than a target threshold value; determining a seed account and a push diffusion mode from the account interaction network; determining candidate accounts from an account interaction network according to a push diffusion mode by taking the seed account as a starting point; and determining a target account used for receiving the target advertisement resource from the candidate accounts by using the second potential influence coefficient and the node influence coefficient matched with the seed account. According to the method and the device, the technical problem of relatively low accuracy of the determined to-be-pushed object in the advertisement resource pushing process in the related technology is solved.

Description

Technical Field

[0001] This application relates to the field of computers, and more specifically, to a method and apparatus for determining a target account, a storage medium, and an electronic device. Background Technology

[0002] To improve the effectiveness of advertising resources, advertising operators often place ads on various application platforms to promote target products, and regard the user groups on these application platforms as potential users of the target products.

[0003] To improve the conversion rate of these potential users to the ads, the most common approach is to rely on the social connections between users within the app platform to determine the ad delivery strategy. In other words, based on the social interaction data generated by each user within the app platform, a target user group similar to the seed users (i.e., the ad audience) is identified from the large number of potential users on the app platform, and then the ads are pushed to this target user group.

[0004] However, even if the target users identified by the above methods have a high degree of operational similarity with the seed users, consistently pushing the same ads to these target users in a uniform manner fails to meet the ever-changing application needs of users. In other words, the methods provided in related technologies for identifying the target accounts for ad pushes are relatively fixed and singular, resulting in the identified target accounts often not being accounts that can be successfully converted, thus causing the technical problem of low push accuracy.

[0005] There is currently no effective solution to the above problems. Summary of the Invention

[0006] This application provides a method and apparatus for determining target accounts, a storage medium, and an electronic device, to at least solve the technical problem of low accuracy in determining the target audience during the advertising resource push process.

[0007] According to one aspect of the embodiments of this application, a method for determining a target account is provided, comprising: constructing an account interaction network using interactive behavior data generated by each object account in an application platform, wherein each node in the account interaction network represents an object account, and each edge in the account interaction network represents an interactive operation between two connected object accounts; dividing the account interaction network to obtain multiple account communities, wherein the association closeness between object accounts contained in the same account community is greater than a target threshold, and each edge is configured with a first potential influence coefficient determined based on the interactive operations between nodes in the account community; and further dividing the account interaction network into multiple account communities. Seed accounts and push diffusion patterns are determined in the dynamic network. Seed accounts are accounts that meet the push conditions that are compatible with the target advertising resources to be pushed. The push diffusion pattern indicates the way to diffuse the target advertising resources based on the potential influence of the seed account on other accounts in the account interaction network. Starting from the seed account, candidate accounts are determined from the account interaction network according to the push diffusion pattern. The first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to a second potential influence coefficient. Using the second potential influence coefficient and the node influence coefficient that matches the seed account, the target account for receiving the target advertising resources is determined from the candidate accounts.

[0008] Optionally, the above-mentioned determination of seed accounts and push diffusion patterns from the account interaction network includes: determining an initial set of seed accounts based on the account attributes of each account in the account interaction network, wherein the account attributes include whether each account meets the push conditions that are compatible with the target advertising resources to be pushed; determining seed accounts that meet preset conditions from the initial set of seed accounts; and determining the push diffusion pattern based on the diffusion ratio, wherein the diffusion ratio is the ratio between the number of target accounts determined from the account interaction network and the number of seed accounts.

[0009] Optionally, the above-mentioned determination of seed accounts that meet the preset conditions from the initial seed account set includes: determining the evaluation factors of each account by using the social capabilities and account activity of each account in the initial seed account set; sorting the evaluation factors of each account in descending order, and determining the top N accounts as seed accounts, where N is a positive integer greater than or equal to 1.

[0010] Optionally, the above-mentioned method of identifying candidate accounts from the account interaction network starting with a seed account and following a push-and-diffusion model includes: determining a predicted potential influence coefficient between nodes in the account community based on a first potential influence coefficient determined by the interaction between nodes in the account community and the interaction between nodes in the account community, wherein the first potential influence coefficient represents the potential influence coefficient determined when there is interaction between nodes, and the predicted potential influence coefficient includes a second potential influence coefficient; and using the predicted potential influence coefficient, candidate accounts are identified from multiple account communities.

[0011] Optionally, the determination of the predicted potential influence coefficient of the edges between nodes in the account community based on the first potential influence coefficient determined by the interaction operations between nodes in the account community and the interaction operations between nodes in the account device includes: performing singular value decomposition on the target adjacency matrix when the diffusion ratio is greater than a preset threshold to obtain a first implicit feature vector and a second implicit feature vector, wherein the target adjacency matrix is ​​a matrix determined based on the first potential influence coefficient determined by the interaction operations between nodes in the account community, the dimension of the target adjacency matrix is ​​M×M, M is the number of all object accounts in the account interaction network, and M is a positive integer greater than or equal to 2; determining the similarity between nodes in the account community based on the first implicit feature vector and the second implicit feature vector; and determining the similarity between nodes as the predicted potential influence coefficient of the edges between nodes.

[0012] Optionally, the above-mentioned determination of the predicted potential influence coefficient between nodes in the account community based on the first potential influence coefficient determined by the interaction operations between nodes in the account community and the interaction operations between nodes in the account community includes: when the diffusion ratio is less than or equal to a preset threshold, determining the predicted potential influence coefficient configured for the edge between each node according to the influence weight corresponding to the interaction operations between each node in the account community.

[0013] Optionally, determining the predicted potential influence coefficient between nodes based on the influence weights corresponding to the interactive operations between nodes within the account community includes: determining the first potential influence coefficient corresponding to the first interactive operation when there is a first interactive operation between the first object account indicated by the first node and the second object account indicated by the second node within the account community; determining the first potential influence coefficient corresponding to the second interactive operation when there is a second interactive operation between the second object account and the third object account indicated by the third node within the account community; and determining the product between the first potential influence coefficient corresponding to the first interactive operation and the first potential influence coefficient corresponding to the second interactive operation when there is no interactive operation between the first object account and the third object account, and determining the product as the predicted potential influence coefficient corresponding to the edge formed by connecting the first object account and the third object account.

[0014] Optionally, the above-mentioned method of determining the target account for receiving the target advertising resources from the candidate accounts using the second potential influence coefficient and the node influence coefficient matching the seed account includes: when the number of seed accounts is F and the number of candidate accounts is K, determining the target influence coefficient of each candidate account among the K candidate accounts being influenced by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account, where F and K are positive integers greater than or equal to 2; sorting the target influence coefficients of each candidate account among the K candidate accounts, and determining the Q candidate accounts corresponding to the top Q target influence coefficients as the target accounts for receiving the target advertising resources, where Q is a positive integer greater than or equal to 1 and less than or equal to K.

[0015] Optionally, the above-mentioned determination of the target influence coefficient of each candidate account among the K candidate accounts being influenced by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account includes: determining the i-th target influence coefficient of the i-th candidate account being influenced by the F seed accounts in the following way, where i is a positive integer greater than or equal to 1 and less than or equal to K: determining the influence coefficient of the i-th candidate account being influenced by each of the F seed accounts, obtaining F influence coefficients; and determining the influence coefficient with the largest value among the F influence coefficients as the target influence coefficient of the i-th candidate account being influenced by the F seed accounts.

[0016] Optionally, determining the influence coefficient of the i-th candidate account by each of the F seed accounts includes: determining the j-th influence coefficient of the i-th candidate account by the j-th seed account through the following steps, where j is a positive integer greater than or equal to 1 and less than or equal to F: determining the j-th second potential influence coefficient configured for the edge between the i-th candidate account and the j-th seed account; and multiplying the j-th node influence coefficient and the j-th second potential influence coefficient that match the j-th seed account to obtain the j-th influence coefficient.

[0017] Optionally, the above-mentioned determination of the target influence coefficient of each candidate account among the K candidate accounts being influenced by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account further includes: determining the i-th target influence coefficient of the i-th candidate account being influenced by the F seed accounts in the following way, where i is a positive integer greater than or equal to 1 and less than or equal to K: determining the influence coefficient of the i-th candidate account being influenced by each of the F seed accounts to obtain the F influence coefficients; and performing a weighted summation of the F influence coefficients to obtain the target influence coefficient of the i-th candidate account being influenced by the F seed accounts.

[0018] Optionally, before determining the target account for receiving the target advertising resources from the candidate accounts using the second potential influence coefficient and the node influence coefficient matched with the seed account, the above method further includes: filtering out a group of accounts that do not meet the conditions from the candidate accounts, wherein the conditions include that the accounts in the group are seed accounts and that the accounts in the group are abnormal accounts.

[0019] Optionally, each of the aforementioned edges is configured with a first potential influence coefficient determined based on the interaction operations between nodes within the account community. This includes: when there is a first type of interaction operation and a second type of interaction operation between the first node and the second node within the account community, determining the first product as the interaction frequency of the first type of interaction operation and the first interaction operation weight of the first type of interaction operation; determining the second product as the interaction frequency of the second type of interaction operation and the second interaction operation weight of the second type of interaction operation; and summing the first product and the second product to obtain the first potential influence coefficient configured for the edge between the first node and the second node.

[0020] According to another aspect of the embodiments of this application, a target account determination device is also provided, comprising: a first processing unit, configured to construct an account interaction network using interactive behavior data generated by each object account in an application platform, wherein each node in the account interaction network represents an object account, and each edge in the account interaction network represents an interactive operation between two connected object accounts; a partitioning unit, configured to partition the account interaction network to obtain multiple account communities, wherein the association closeness between object accounts contained in the same account community is greater than a target threshold, and each edge is configured with a first potential influence coefficient determined based on the interactive operations between nodes in the account community; and a second processing unit, configured to... Seed accounts and push diffusion patterns are determined in the account interaction network. Seed accounts are accounts that meet the push conditions that are compatible with the target advertising resources to be pushed. The push diffusion pattern indicates the method of determining the diffusion of target advertising resources based on the potential influence of seed accounts on non-seed accounts in the account interaction network. The third processing unit is used to determine candidate accounts from the account interaction network starting from the seed accounts and according to the push diffusion pattern. The first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to a second potential influence coefficient. The fourth processing unit is used to determine the target account for receiving the target advertising resources from the candidate accounts using the second potential influence coefficient and the node influence coefficient that matches the seed account.

[0021] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, which is used to execute the above-described method for determining the target account when the electronic device is run.

[0022] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0023] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the method for determining the target account through the computer program.

[0024] By constructing an account interaction network based on the interactive behavior data generated by each object account in the application platform, and dividing the account interaction network into multiple account communities, the potential influence coefficient between accounts is calculated within each account community according to the push diffusion mode and the interactive operations between each node. Combining the node influence coefficient of the seed account itself and the potential influence coefficient between accounts, the influence coefficient of candidate accounts on the seed account is calculated. Finally, based on the influence coefficient, the target accounts for receiving target advertising resources are determined from the candidate accounts. In other words, this embodiment of the application accurately calculates the potential influence coefficient between nodes by combining the interactive operations between object accounts, thereby determining the target accounts that are significantly influenced by the interactive behavior of the seed account, improving the accuracy of the determined target accounts for receiving target advertising resources, and increasing the advertising conversion rate. Attached Figure Description

[0025] The accompanying drawings, which are provided to further understand this application and constitute a part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application.

[0026] Figure 1 This is a schematic diagram illustrating an application scenario of an optional method for determining a target account according to an embodiment of this application;

[0027] Figure 2 This is a flowchart of an optional method for determining a target account according to an embodiment of this application;

[0028] Figure 3 This is an overall schematic diagram of an optional method for determining a target account according to an embodiment of this application;

[0029] Figure 4 This is a diagram illustrating the process of determining the first potential influence coefficient between accounts based on interaction type and interaction density.

[0030] Figure 5 This is a product-side schematic diagram of an optional method for determining a target account according to an embodiment of this application;

[0031] Figure 6 This is an optional method for selecting a seed account according to an embodiment of this application;

[0032] Figure 7 This is a schematic diagram illustrating an optional process for determining a target account according to an embodiment of this application;

[0033] Figure 8 This is a schematic diagram illustrating another optional process for determining the target account according to an embodiment of this application;

[0034] Figure 9A flowchart of an optional method for determining a target account according to an embodiment of this application;

[0035] Figure 10 This is a schematic diagram of the structure of an optional target account determination device according to an embodiment of this application;

[0036] Figure 11 This is a schematic diagram of the structure of an optional electronic device according to an embodiment of this application. Detailed Implementation

[0037] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0038] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0039] The technical solutions in this application will comply with legal regulations during implementation. When performing operations according to the technical solutions in the embodiments, the data used will not involve user privacy, ensuring that the operation process is compliant and legal while guaranteeing data security.

[0040] In addition, when the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant regulations and standards of the relevant countries or regions.

[0041] According to one aspect of the embodiments of this application, a method for determining a target account is provided. As an optional implementation, the above-described method for determining a target account may be applied, but is not limited to, to applications such as... Figure 1 The application scenarios shown are as follows. In, for example... Figure 1In the application scenario shown, the target terminal 102 can communicate with the server 106 via network 104, but is not limited to this. The server 106 can perform operations on the database 108, such as write or read data operations. The target terminal 102 may include, but is not limited to, a human-computer interaction screen, a processor, and a memory. The human-computer interaction screen may be used to display the account operation interface of the application platform on the target terminal 102, the determined target account, etc. The processor may be used to respond to the human-computer interaction operation, execute the corresponding operation, or generate the corresponding instruction and send the generated instruction to the server 106. The memory is used to store relevant processing data, such as the first potential influence coefficient, the push diffusion mode, and the target account, etc.

[0042] Optionally, in this embodiment, the target terminal can be a terminal configured with a target client, which may include, but is not limited to, at least one of the following: mobile phone (such as Android phone, iOS phone, etc.), laptop computer, tablet computer, PDA, MID (Mobile Internet Devices), PAD, desktop computer, smart TV, etc. The target client may be a video client, instant messaging client, browser client, educational client, etc. The network may include, but is not limited to, wired network and wireless network, wherein the wired network includes: local area network, metropolitan area network and wide area network, and the wireless network includes: Bluetooth, WIFI and other networks that enable wireless communication. The server may be a single server, a server cluster composed of multiple servers, or a cloud server.

[0043] The technical solutions in this application embodiment can be applied to scenarios such as advertising resource push, such as game marketing and product advertising push. The main approach is to accurately calculate the relationship strength between each user based on the interaction between candidate users of the target advertising resource, with the seed user as the center, and then use the influence coefficient of the seed user on each candidate user as a reference to select the target users to receive the target advertising resource.

[0044] To address the issue of low accuracy in identifying target audiences during ad delivery, this application proposes a method for determining target accounts. Figure 2 This is a flowchart of a method for determining a target account according to an embodiment of this application. The process includes the following steps S202 to S210.

[0045] It should be noted that the method for determining the target account shown in steps S202 to S210 can be, but is not limited to, performed by an electronic device. The electronic device can be, but is not limited to, a device that... Figure 1 The target terminal or server shown.

[0046] Step S202: Construct an account interaction network using the interaction behavior data generated by each object account in the application platform. Each node in the account interaction network represents an object account, and each edge in the account interaction network represents an interaction operation between two connected object accounts.

[0047] Step S204: Divide the account interaction network to obtain multiple account communities. Among them, the correlation between object accounts contained in the same account community is greater than the target threshold. Each edge is configured with a first potential influence coefficient determined based on the interaction operation between each node in the account community.

[0048] Step S206: Determine the seed account and push diffusion mode from the account interaction network. The seed account is an account that meets the push conditions that are compatible with the target advertising resource to be pushed. The push diffusion mode means determining the way to spread the target advertising resource based on the potential influence of the seed account on other accounts in the account interaction network.

[0049] Step S208: Starting from the seed account, candidate accounts are determined from the account interaction network according to the push diffusion mode, wherein the first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to the second potential influence coefficient.

[0050] Step S210: Using the second potential influence coefficient and the node influence coefficient matching the seed account, determine the target account from the candidate accounts to receive the target advertising resources.

[0051] To facilitate understanding of the technical solutions in the embodiments of this application, the following explanations will use the promotion and marketing of game applications as examples to illustrate the method for determining the target account.

[0052] After a game is launched, it is usually showcased to a select group of users through promotional activities. Some accounts will then register and start using the game. These early users can be referred to as seed accounts, which can help the game spread rapidly and expand its user base.

[0053] In the process of marketing games through seed accounts, the first step is to define the target audience, such as users aged 20-30, or users who spend more than 2 hours online between 20:00 and 24:00.

[0054] The objects to be pushed according to the above conditions are used as the respective object accounts in the application platform in this application embodiment, for example, such as Figure 3As shown, assume that the object accounts in the application platform include u1 to u2. 10 There are a total of 10 accounts. An account interaction network is constructed through the interaction operations between the 10 accounts. The edge between u1 and u2 is used to indicate that there is an interaction operation between accounts u1 and u2, while there is no interaction operation between u1 and u3.

[0055] In the construction of an account interaction network, a first potential influence coefficient is typically assigned to the edge between two accounts that have genuine interaction. The calculation method for this first potential influence coefficient includes:

[0056] Based on the type and density of interaction between the two accounts, determine the weight of each interaction behavior on the strength of the relationship between the two accounts;

[0057] The product of the weight of each interaction behavior on the strength of the relationship between the two accounts and the interaction density of the current interaction behavior is determined as the sub-potential influence coefficient of the current interaction behavior on the interaction behavior between the two accounts.

[0058] The sub-potential influence coefficients corresponding to each interactive behavior are summed to obtain the first potential influence coefficient between the two accounts.

[0059] Specifically, when there is a first type of interaction and a second type of interaction between the first node and the second node in the account community, the product of the interaction frequency of the first type of interaction and the first interaction weight of the first type of interaction is determined as the first product.

[0060] The second product is determined by multiplying the frequency of the second type of interactive operation by the weight of the second type of interactive operation.

[0061] Summing the first and second products yields the first potential influence coefficient assigned to the edge between the first and second nodes.

[0062] In such Figure 4 In one specific embodiment shown, it is assumed that there is a real interaction between accounts u1 and u2, which includes interaction behavior 1 and interaction behavior 2. Then, the correlation strength between accounts u1 and u2 can be calculated based on the type and density of the two interaction behaviors, which can also be understood as the first potential influence coefficient between accounts u1 and u2.

[0063] Specifically, when interaction behavior 1 is sending flowers and interaction behavior 2 is forming a team, the weight of sending flowers on the relationship strength between accounts u1 and u2 is determined to be 0.3, and the weight of forming a team on the relationship strength between accounts u1 and u2 is determined to be 0.7. Assuming that the interaction density of sending flowers is 3 times a day and the activity density of forming a team is 5 times a day, then the total edge weight of the relationship between u1 and u2 is 3×0.3+5×0.7, that is, the first potential influence coefficient between u1 and u2 is 3×0.3+5×0.7.

[0064] By calculating the first potential influence coefficient (edge ​​weight) between two accounts that have interacted, the strength of the association between nodes in the account interaction network can be measured, with the aim of revealing the potential relationships and influence paths between users.

[0065] After determining the first potential influence coefficient between accounts, the corresponding push diffusion mode is selected based on the expected promotion effect and the expected marketing volume. The push diffusion mode is mainly determined based on the diffusion ratio, which will be explained in the following specific examples.

[0066] To avoid calculating the edge weights (which can also be understood as potential influence coefficients) between any two accounts in the entire account interaction network, this embodiment first divides the entire account interaction network into multiple account communities. Accounts with close connections are grouped into the same account community, while accounts with almost no connection are grouped into different account communities. For example... Figure 3 As shown, u1 to u3 are assigned to account community 1, and u4 to u 10 Accounts have been assigned to account community 2, including accounts u1 and u4~u 10 There is almost no connection between them, but there is a close connection between accounts u1 to u3.

[0067] After determining the push notification diffusion model, within each account community, starting with a seed account, candidate accounts are selected from each account community according to the selected push notification diffusion model. During this process, the first potential influence coefficient between accounts that are classified into the same account community and have genuine interactive operations will be updated to the second potential influence coefficient. For example... Figure 3 As shown, the first potential influence coefficient w1 between u1 and u2 is updated to the second potential influence coefficient w'1, and the first potential influence coefficient w2 between u2 and u3 is updated to the second potential influence coefficient w'2. Simultaneously, based on the second potential influence coefficient w'1 between u1 and u2 and the second potential influence coefficient w'2 between u2 and u3, the second potential influence coefficient w'3 between accounts u1 and u3 that have not engaged in any real interaction is predicted.

[0068] Assuming account u1 is the seed account, and the candidate accounts within account community 1 are u2 and u3, then calculate the influence coefficient of candidate accounts u2 and u3 on the seed account u1. By comparing the influence coefficients, determine the target accounts within account community 1 that meet the criteria. Referring to the method for determining target accounts within account community 1, determine the target accounts within other account communities that meet the criteria, and use all determined target accounts as the target accounts.

[0069] In such Figure 5 In one specific embodiment shown, the target account is determined through the following steps:

[0070] S51, Identify the target seed account;

[0071] Suppose that the target seed accounts for this marketing campaign (e.g., users who participated in the previous promotion and subsequently converted) are entered into the marketing system interface on the terminal device. Seed accounts can refer to, but are not limited to, the most core early users of the product. They usually have high activity and influence and can help the product or service spread quickly and expand the user base.

[0072] It's easy to understand that for a game, the number of early adopters (seed accounts) is far greater than 200. Therefore, a predetermined number of target seed accounts can be selected from the existing seed accounts, but not limited to the following methods:

[0073] Based on the account attributes of each account in the account interaction network, an initial set of seed accounts is determined. The account attributes include whether each account meets the push conditions that are compatible with the target advertising resources to be pushed.

[0074] Seed accounts that meet the preset conditions are identified from the initial set of seed accounts.

[0075] The push diffusion mode is determined based on the diffusion ratio, where the diffusion ratio is the ratio between the number of target accounts identified in the account interaction network and the number of seed accounts.

[0076] Assuming there are 1000 accounts in the account interaction network, based on whether they have registered and used the current game application, 300 seed accounts that have already used the current game in the early stages are identified from the 1000 accounts. Then, 200 target seed accounts are selected from these 300 initial seed accounts to meet the needs of this marketing campaign.

[0077] As an alternative example, methods for selecting target seed accounts from the initial set of seed accounts include:

[0078] By utilizing the social capabilities and account activity levels of each account in the initial seed account set, the evaluation factors for each account are determined.

[0079] The evaluation factors of each account are sorted in descending order, and the top N accounts are identified as seed accounts.

[0080] like Figure 6 As shown, assuming that each account's evaluation factors e1 to e2 are calculated based on the amount spent, social skills, and account activity level during the use of the game application in the current marketing campaign, the evaluation factors for each account are calculated sequentially. 10 By ranking the various evaluation factors, the top N seed accounts are selected as the target seed accounts for this marketing campaign, according to the required number.

[0081] By using the above methods, target seed accounts with high social skills and account activity can be selected from existing seed accounts, thereby enhancing the potential influence of each seed account on the signatures of candidate accounts within the account interaction network.

[0082] S52, set the expected amount of marketing data to be sent;

[0083] This means setting the number of recipients to be targeted (which can also be understood as the number of target accounts) based on the expected marketing results. For example, a total of 10,000 campaign ads may be sent.

[0084] S53, the system automatically calculates the diffusion ratio;

[0085] The diffusion ratio can be, but is not limited to, the ratio between the number of target accounts and the number of seed accounts. For example, if the number of target accounts is 10,000 and the number of seed accounts is 200, then the diffusion ratio is the ratio of 10,000 to 200, which is 50.

[0086] Diffusion ratio can be used, but is not limited to, to measure the effectiveness of advertising. When combined with factors such as advertising costs, advertising operators may expect different promotional effects based on the actual situation. Therefore, the value of diffusion ratio is not fixed, but can be adjusted.

[0087] By adaptively adjusting the diffusion ratio, this method solves the problem that it can only be used when there are a relatively large number of seed accounts and the account characteristics are relatively clustered, compared to related technologies. It also solves the problem that the methods in related technologies are not suitable for relatively sparse seed accounts (few in number or diverse in type) and cannot utilize users' social network information, which causes a great deal of resource waste for the operation of Internet products.

[0088] It is evident that by adjusting the diffusion ratio, the method of determining target accounts based on social network information in this application embodiment can be made more inclusive, thereby improving the practicality of the method.

[0089] S54, the system automatically outputs a preset number of target accounts.

[0090] For example, based on the diffusion ratio and the previous interaction behavior of 200 seed accounts in the game, the system automatically outputs 10,000 accounts and sends these 10,000 accounts to the target advertising resources.

[0091] By constructing an account interaction network based on the interactive behavior data generated by each object account in the application platform, and dividing the account interaction network into multiple account communities, the potential influence coefficient between accounts is calculated within each account community according to the push diffusion mode and the interactive operations between each node. Combining the node influence coefficient of the seed account itself and the potential influence coefficient between accounts, the influence coefficient of candidate accounts on the seed account is calculated. Finally, based on the influence coefficient, the target accounts for receiving target advertising resources are determined from the candidate accounts. In other words, this embodiment of the application accurately calculates the potential influence coefficient between nodes by combining the interactive operations between object accounts, thereby determining the target accounts that are significantly influenced by the interactive behavior of the seed account, improving the accuracy of the determined target accounts for receiving target advertising resources, and increasing the advertising conversion rate.

[0092] As an optional example, the above method, starting with a seed account, identifies candidate accounts from the account interaction network according to a push-and-diffusion pattern, including:

[0093] Based on the first potential influence coefficient determined by the interaction between various nodes in the account community and the interaction between various nodes in the account community, the predicted potential influence coefficient between various nodes in the account community is determined. The first potential influence coefficient represents the potential influence coefficient determined when there is interaction between various nodes, and the predicted potential influence coefficient includes the second potential influence coefficient.

[0094] Candidate accounts were identified from multiple account communities by using the predicted potential influence coefficient.

[0095] As can be seen from the description in the above embodiments, in the account interaction network, there are real interactive operations between some accounts. However, after the community is divided, within the same account community, affected by the interactive operations between other accounts, there may be a predictable potential influence between accounts that originally had no real interactive operations within the same account community.

[0096] For example, such as Figure 3As shown, no real interaction occurred between accounts u1 and u3, but by considering the correlation between u1 and u2, and u2 and u3, the predicted influence coefficient between accounts u1 and u3 can be determined.

[0097] At the same time, after the community division, the first potential influence coefficient between u1 and u2 will be updated to the second potential influence coefficient, and the first potential influence coefficient between u2 and u3 will also be updated to the second potential influence coefficient.

[0098] By using the predicted potential influence coefficient, candidate accounts are identified from multiple account communities. For example, candidate accounts u6 to u6 are identified from account community 2. 10 .

[0099] As an optional implementation, the predicted potential influence coefficients configured for the edges between nodes within the account community, determined based on the first potential influence coefficient determined by the interaction operations between nodes within the account community and the interaction operations between nodes within the account device, include:

[0100] When the diffusion ratio is greater than a preset threshold, singular value decomposition is performed on the target adjacency matrix to obtain the first implicit eigenvector and the second implicit eigenvector. The target adjacency matrix is ​​a matrix determined based on the first potential influence coefficient determined by the interaction operation between various nodes in the account community. The dimension of the target adjacency matrix is ​​M×M, where M is the number of all object accounts in the account interaction network and M is a positive integer greater than or equal to 2.

[0101] Based on the first and second implicit feature vectors, the similarity between nodes within the account community is determined.

[0102] The similarity between nodes is determined as the predicted potential influence coefficient configured for the edges between nodes.

[0103] Assuming a marketing campaign has a diffusion ratio of 150, such as Figure 7 As shown, since the diffusion ratio is greater than the preset threshold of 100, it also means that the number of seed accounts is very small relative to the number of target accounts. Therefore, the first push diffusion mode will be adopted, which is to use the similarity calculation method to calculate the predicted potential influence coefficient between accounts in the account community (which can also be understood as the predicted edge weight configured between accounts).

[0104] Before judging the diffusion ratio, an M×M matrix is ​​constructed based on the first potential influence coefficient between accounts before the account community is divided. Then, the target adjacency matrix is ​​obtained by normalizing the M×M matrix, where M is the number of all object accounts in the account interaction network.

[0105] By performing singular value decomposition on the target adjacency matrix, we obtain the first implicit eigenvector P and the second implicit eigenvector Q. The first implicit eigenvector P has a dimension of M×i, and the second implicit eigenvector Q has a dimension of i×M, where i takes a relatively small value.

[0106] Based on the first implicit feature vector P and the second implicit feature vector Q, the similarity between each pair of accounts within the account community is calculated, and this similarity is determined as the predicted potential influence coefficient configured for the edges between each node.

[0107] As another optional example, the predicted potential influence coefficient between nodes in the account community is determined based on the first potential influence coefficient determined by the interaction between nodes within the account community, and the interaction between nodes within the account community. This includes:

[0108] When the diffusion ratio is less than or equal to a preset threshold, the predicted potential influence coefficient of the edges between each node is determined based on the influence weights corresponding to the interactive operations between each node in the account community.

[0109] For example, in another marketing campaign with a diffusion ratio of 50, such as Figure 7 As shown, since the diffusion ratio is less than the preset threshold of 100, which means that the number of seed accounts is very large relative to the number of target accounts, the second push diffusion mode will be adopted. This mode calculates the predicted potential influence coefficient between each pair of accounts within the account community using the correlation of influence (which can also be understood as the predicted edge weights configured between accounts). Specifically, this includes:

[0110] In the case where there is a first interactive operation between the first target account indicated by the first node and the second target account indicated by the second node within the account community, determine the first potential influence coefficient corresponding to the first interactive operation;

[0111] In the event that there is a second interactive operation between the second object account and the third object account indicated by the third node within the account community, determine the first potential influence coefficient corresponding to the second interactive operation;

[0112] When there is no interaction between the first object account and the third object account, determine the product between the first potential influence coefficient corresponding to the first interaction and the first potential influence coefficient corresponding to the second interaction, and determine the product as the predicted potential influence coefficient corresponding to the edge formed by connecting the first object account and the third object account.

[0113] For example, assuming there is a flower-sending behavior between accounts u1 and u2, then the first potential influence coefficient w1 corresponding to the flower-sending behavior is determined; if there is a team-up behavior between accounts u2 and u3, then the first potential influence coefficient w2 corresponding to the team-up behavior is determined.

[0114] In the absence of any interaction between accounts u1 and u3, the predicted potential influence coefficient of the edge formed by connecting accounts u1 and u3 is calculated as w1*w2 using the prediction edge weight = first-degree influence * second-degree influence * (for example, if there is a third-degree edge weight).

[0115] Using the above method, the influence coefficient between any two accounts within an account community can be directly predicted when there are a large number of seed accounts and a small number of target accounts.

[0116] As an optional example, the above method of using a second potential influence coefficient and a node influence coefficient matching the seed account to determine the target account from the candidate accounts for receiving the target advertising resources includes:

[0117] With F seed accounts and K candidate accounts, the target influence coefficient of each candidate account among the K candidate accounts is determined by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account, where F and K are positive integers greater than or equal to 2.

[0118] Sort the target influence coefficients of each of the K candidate accounts, and determine the Q candidate accounts corresponding to the top Q target influence coefficients as the target accounts for receiving target advertising resources, where Q is a positive integer greater than or equal to 1 and less than or equal to K.

[0119] In this embodiment of the application, the target account is determined mainly based on the potential influence of the seed account on the candidate account. Specifically, the potential influence coefficient of the seed account on the candidate account = π node influence coefficient * edge weight. The node influence coefficient (PageRank) is an algorithm used to evaluate the importance of a webpage. Its basic idea is to measure the link relationship between nodes and determine the weight of the node by the number and quality of the links.

[0120] In this embodiment of the application, the node influence coefficient is used to represent the degree of influence of the seed account on other accounts in the account interaction network. It is mainly used to measure the possibility of other accounts registering and using game applications in advertising resources under the influence of the seed account. The edge weight is the second potential influence coefficient of the edge configuration formed by connecting the seed account and the candidate account.

[0121] By using the second potential influence coefficient and the node influence coefficient corresponding to the seed account, the influence coefficient of each candidate account is determined, thereby identifying the target account.

[0122] As an optional implementation, the above method determines the target influence coefficient of each of the K candidate accounts by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account, including:

[0123] The influence coefficient of the i-th target affected by the F seed accounts on the i-th candidate account is determined as follows, where i is a positive integer greater than or equal to 1 and less than or equal to K:

[0124] Determine the influence coefficient of the i-th candidate account on each of the F seed accounts, and obtain the F influence coefficients;

[0125] The largest of the F affected coefficients is determined as the target affected coefficient of the i-th candidate account being affected by the F seed accounts.

[0126] like Figure 7 As shown, assuming there are seed accounts u1, u2 and u3 in an account community, and candidate accounts u4 to u8, calculate the influence coefficient of the three seed accounts on each candidate account to obtain the influence coefficient of each candidate account.

[0127] For example, the influence coefficient of seed account 1 on candidate account u4 is w. 14 The influence coefficient of seed account 2 on candidate account u4 is w. 24 The influence coefficient of seed account 3 on candidate account u4 is w. 34 Then, the total influence coefficient of candidate account 1 = MAX(the influence coefficient of the seed user on this user), assuming w 14 >w 24 >w 34 Therefore, the total influence coefficient of candidate account u4 is determined to be w. 14 .

[0128] Following the same method, the influence coefficients of seed accounts 1 through 3 on candidate account u5 were determined as w, respectively. 15 w 25 w 35 Assume w 15 >w 25 >w 35 Therefore, the total influence coefficient of candidate account u5 is determined to be w. 15 Similarly, the total influence coefficients of candidate accounts u6 to u8 are obtained respectively.

[0129] Then, the total influence coefficient of each candidate account is compared, and the t candidate accounts corresponding to the top t total influence coefficients with the largest total influence coefficients are identified as target accounts within that account's community. Similarly, target accounts within other account communities can be identified sequentially.

[0130] As an optional example, the determination of the influence coefficient of the i-th candidate account by each of the F seed accounts includes:

[0131] The following steps are used to determine the j-th influence coefficient of the i-th candidate account being affected by the j-th seed account, where j is a positive integer greater than or equal to 1 and less than or equal to F:

[0132] The j-th second potential influence coefficient is determined as the edge configured between the i-th candidate account and the j-th seed account;

[0133] Multiply the influence coefficient of the jth node and the second potential influence coefficient that match the jth seed account to obtain the jth influenced coefficient.

[0134] Specifically, the potential influence coefficient of the seed account on the candidate account = ∏ node influence coefficient * edge weight, where the node influence coefficient is used to represent the degree of influence of the seed account on other accounts in the account interaction network. It is mainly used to measure the likelihood of other accounts registering and using game applications in advertising resources under the influence of the seed account. The edge weight is the second potential influence coefficient of the edge configuration formed by connecting the seed account and the candidate account.

[0135] As another optional example, the above method of determining the target influence coefficient of each of the K candidate accounts by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account also includes:

[0136] The influence coefficient of the i-th target affected by the F seed accounts on the i-th candidate account is determined as follows, where i is a positive integer greater than or equal to 1 and less than or equal to K:

[0137] Determine the influence coefficient of the i-th candidate account on each of the F seed accounts, and obtain the F influence coefficients;

[0138] The target influence coefficient of the i-th candidate account is obtained by weighted summation of the F influence coefficients and the F seed accounts.

[0139] like Figure 8As shown, assuming there are seed accounts u1, u2 and u3 in an account community, and candidate accounts u4 to u8, calculate the influence coefficient of the three seed accounts on each candidate account to obtain the influence coefficient of each candidate account.

[0140] For example, the influence coefficient of seed account 1 on candidate account u4 is w. 14 The influence coefficient of seed account 2 on candidate account u4 is w. 24 The influence coefficient of seed account 3 on candidate account u4 is w. 34 Then the total influence coefficient of candidate account 1, W4 = w 14 +w 24 +w 34 .

[0141] Following the same method, the influence coefficients of seed accounts 1 through 3 on candidate account u5 were determined as w, respectively. 15 w 25 w 35 Then the total influence coefficient of candidate account 1, W5 = w 15 +w 254 +w 35 Similarly, the total influence coefficients of candidate accounts u6 to u8 are obtained respectively.

[0142] Obviously, the above method of directly summing the influence coefficients of each seed account on the candidate account u4 is just one example and is not limited to it. For example, it could also be a weighted summation based on different weights.

[0143] Then, the total influence coefficient of each candidate account is compared, and the t candidate accounts corresponding to the top t total influence coefficients with the largest total influence coefficients are identified as target accounts within that account's community. Similarly, target accounts within other account communities can be identified sequentially.

[0144] By using the weighted summation method described above, the potential influence of seed accounts on candidate accounts can be determined more objectively. This results in a higher conversion rate of the target accounts identified from the candidate accounts for the target advertising resources, achieving the technical effect of improving marketing effectiveness.

[0145] As an alternative example, before determining the target account for receiving the target advertising resources from the candidate accounts using the second potential influence coefficient and the node influence coefficient matched with the seed account, the above method further includes:

[0146] Select a group of accounts from the candidate accounts that do not meet the criteria, including whether the accounts in the group are seed accounts or abnormal accounts.

[0147] After determining the total impact coefficient of candidate accounts according to the above embodiments, it is also necessary to conduct preliminary screening of candidate accounts based on business needs and rules. For example, filtering out accounts that have already been converted and do not need to be promoted again, and filtering out known abnormal accounts.

[0148] Then, the filtered candidate accounts are sorted according to the total impact coefficient, and the top N candidate accounts are selected as target accounts, where N is the preset number of accounts that need to send push messages.

[0149] To more clearly understand the method for determining the target account mentioned above, the following will combine... Figure 9 The overall flowchart shown further describes it.

[0150] S902, based on the interaction operations between the two object accounts connected by each edge in the account interaction network, determine the first potential influence coefficient between each node, and determine the adjacency matrix based on the first potential influence coefficient;

[0151] Specifically, an account interaction network is constructed based on the interaction behavior data generated by each object account in the account interaction network, and a first potential influence coefficient is determined based on the interaction type and interaction density between accounts. For details, please refer to the descriptions in some of the above embodiments, which will not be repeated here.

[0152] S904, calculates the influence of seed accounts;

[0153] S906, Determine whether the diffusion ratio is greater than a preset threshold;

[0154] If the value is greater than the preset threshold, proceed to step S908; otherwise, proceed to step S912.

[0155] S908, perform singular value decomposition on the adjacency matrix to generate the first implicit eigenvector and the second implicit eigenvector;

[0156] The adjacency matrix is ​​decomposed using SVD++, which is an extension of Singular Value Decomposition.

[0157] S910, Based on the first and second implicit feature vectors, calculate the implicit factor similarity between accounts to obtain the predicted edge weights between accounts;

[0158] Among them, the prediction edge weight can be understood as the prediction potential influence coefficient between accounts.

[0159] S912, influence is transmitted according to degree;

[0160] S914 determines the predicted potential influence coefficient based on the first-degree influence, second-degree influence, etc. between accounts;

[0161] Specifically, the predicted potential influence coefficient is equal to the product of the potential influence coefficients of the edges containing the multiple sub-paths.

[0162] For example, if there is a first interactive operation between a first target account indicated by a first node in the account community and a second target account indicated by a second node, a first potential influence coefficient corresponding to the first interactive operation is determined; if there is a second interactive operation between a second target account and a third target account indicated by a third node in the account community, a first potential influence coefficient corresponding to the second interactive operation is determined.

[0163] In the absence of any interaction between the first object account and the third object account, the product between the first potential influence coefficient corresponding to the first interaction and the first potential influence coefficient corresponding to the second interaction is determined, and the product is determined as the predicted potential influence coefficient corresponding to the edge formed by connecting the first object account and the third object account.

[0164] S916, calculate the total impact coefficient for each candidate account, and determine the target account based on the total impact coefficient.

[0165] There are two methods, please refer to the details. Figure 7 and Figure 8 The details of the previous descriptions will not be repeated here.

[0166] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0167] According to another aspect of the embodiments of this application, as follows is also provided Figure 10 A target account determination device is shown, the device comprising:

[0168] The first processing unit 1002 is used to construct an account interaction network by utilizing the interaction behavior data generated by each object account in the application platform. Each node in the account interaction network is used to represent an object account, and each edge in the account interaction network is used to represent the interaction operation between two connected object accounts.

[0169] The partitioning unit 1004 is used to partition the account interaction network to obtain multiple account communities. Among them, the correlation between object accounts contained in the same account community is greater than the target threshold. Each edge is configured with a first potential influence coefficient determined based on the interaction operation between each node in the account community.

[0170] The second processing unit 1006 is used to determine the seed account and the push diffusion mode from the account interaction network. The seed account is an account that meets the push conditions that are compatible with the target advertising resource to be pushed. The push diffusion mode means determining the way to diffuse the target advertising resource based on the potential influence of the seed account on other accounts in the account interaction network.

[0171] The third processing unit 1008 is used to determine candidate accounts from the account interaction network starting from the seed account and according to the push diffusion mode, wherein the first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to the second potential influence coefficient.

[0172] The fourth processing unit 1010 is used to determine the target account for receiving the target advertising resources from the candidate accounts by using the second potential influence coefficient and the node influence coefficient matched with the seed account.

[0173] Optionally, the second processing unit 1006 includes:

[0174] The first processing module is used to determine the initial seed account set based on the account attributes of each account in the account interaction network. The account attributes include whether each account meets the push conditions that are compatible with the target advertising resources to be pushed.

[0175] The second processing module is used to determine the seed accounts that meet the preset conditions from the initial set of seed accounts;

[0176] The third processing module is used to determine the push diffusion mode based on the diffusion ratio, where the diffusion ratio is the ratio between the number of target accounts determined from the account interaction network and the number of seed accounts.

[0177] Optionally, the second processing module mentioned above includes:

[0178] The first processing submodule is used to determine the evaluation factors for each account by utilizing the social capabilities and account activity of each account in the initial seed account set.

[0179] The second processing submodule is used to sort the evaluation factors of each account in descending order and determine the top N accounts as seed accounts.

[0180] Optionally, the third processing unit 1008 includes:

[0181] The fourth processing module is used to determine the predicted potential influence coefficient between nodes in the account community based on the first potential influence coefficient determined by the interaction between nodes in the account community and the interaction between nodes in the account community. The first potential influence coefficient represents the potential influence coefficient determined when there is interaction between nodes, and the predicted potential influence coefficient includes the second potential influence coefficient.

[0182] The fifth processing module is used to identify candidate accounts from multiple account communities by using the predicted potential influence coefficient.

[0183] Optionally, the fourth processing module mentioned above includes:

[0184] The third processing submodule is used to perform singular value decomposition on the target adjacency matrix when the diffusion ratio is greater than a preset threshold, to obtain the first implicit feature vector and the second implicit feature vector. The target adjacency matrix is ​​a matrix determined based on the first potential influence coefficient determined by the interaction operation between various nodes in the account community. The dimension of the target adjacency matrix is ​​M×M, where M is the number of all object accounts in the account interaction network and M is a positive integer greater than or equal to 2.

[0185] The fourth processing submodule is used to determine the similarity between nodes within the account community based on the first implicit feature vector and the second implicit feature vector.

[0186] The fifth processing submodule is used to determine the similarity between nodes as the predicted potential influence coefficient configured by the edges between the nodes.

[0187] Optionally, the fourth processing module mentioned above includes:

[0188] The sixth processing submodule is used to determine the predicted potential influence coefficient of the edges between each node based on the influence weights corresponding to the interactive operations between each node in the account community when the diffusion ratio is less than or equal to a preset threshold.

[0189] Optionally, the fourth processing module mentioned above further includes:

[0190] The seventh processing submodule is used to determine the first potential influence coefficient corresponding to the first interactive operation when there is a first interactive operation between the first object account indicated by the first node and the second object account indicated by the second node in the account community.

[0191] The eighth processing submodule is used to determine the first potential influence coefficient corresponding to the second interactive operation when there is a second interactive operation between the second object account and the third object account indicated by the third node in the account community.

[0192] The ninth processing submodule is used to determine the product between the first potential influence coefficient corresponding to the first interaction operation and the first potential influence coefficient corresponding to the second interaction operation when there is no interaction operation between the first object account and the third object account, and to determine the product as the predicted potential influence coefficient corresponding to the edge formed by connecting the first object account and the third object account.

[0193] Optionally, the fourth processing unit 1010 includes:

[0194] The sixth processing module is used to determine the target influence coefficient of each candidate account among the K candidate accounts by the F seed accounts, based on the second potential influence coefficient and the node influence coefficient matching the seed account, when the number of seed accounts is F and the number of candidate accounts is K, where F and K are positive integers greater than or equal to 2.

[0195] The sorting module is used to sort the target influence coefficient of each candidate account among the K candidate accounts, and determine the Q candidate accounts corresponding to the top Q target influence coefficients as the target accounts for receiving target advertising resources, where Q is a positive integer greater than or equal to 1 and less than or equal to K.

[0196] Optionally, the sixth processing module mentioned above includes:

[0197] The tenth processing submodule is used to determine the target influence coefficient of the i-th candidate account being affected by F seed accounts in the following way, where i is a positive integer greater than or equal to 1 and less than or equal to K:

[0198] Determine the influence coefficient of the i-th candidate account on each of the F seed accounts, and obtain the F influence coefficients;

[0199] The largest of the F affected coefficients is determined as the target affected coefficient of the i-th candidate account being affected by the F seed accounts.

[0200] Optionally, the sixth processing module mentioned above includes:

[0201] The eleventh processing submodule is used to determine the j-th influence coefficient of the i-th candidate account being affected by the j-th seed account through the following steps, where j is a positive integer greater than or equal to 1 and less than or equal to F:

[0202] The j-th second potential influence coefficient is determined as the edge configured between the i-th candidate account and the j-th seed account;

[0203] Multiply the influence coefficient of the jth node and the second potential influence coefficient that match the jth seed account to obtain the jth influenced coefficient.

[0204] Optionally, the sixth processing module mentioned above includes:

[0205] The twelfth processing submodule is used to determine the target influence coefficient of the i-th candidate account being affected by F seed accounts in the following way: where i is a positive integer greater than or equal to 1 and less than or equal to K:

[0206] Determine the influence coefficient of the i-th candidate account on each of the F seed accounts, and obtain the F influence coefficients;

[0207] The target influence coefficient of the i-th candidate account is obtained by weighted summation of the F influence coefficients and the F seed accounts.

[0208] Optionally, the above-mentioned device further includes:

[0209] The filtering unit is used to filter out a group of accounts that do not meet the conditions before determining the target account for receiving the target advertising resources from the candidate accounts by using the second potential influence coefficient and the node influence coefficient matched with the seed account. The conditions include that the accounts in the group are seed accounts and that the accounts in the group are abnormal accounts.

[0210] Optionally, the above-mentioned dividing unit 1004 includes:

[0211] The seventh processing module is used to determine the first product as the product between the interaction frequency of the first type of interaction and the first interaction weight of the first type of interaction when there is a first type of interaction and a second type of interaction between the first node and the second node in the account community.

[0212] The eighth processing module is used to determine the second product as the product between the interaction frequency of the second type of interactive operation and the weight of the second type of interactive operation in the second interactive operation.

[0213] The ninth processing module is used to sum the first product and the second product to obtain the first potential influence coefficient configured for the edge between the first node and the second node.

[0214] By applying the aforementioned device to construct an account interaction network based on the interactive behavior data generated by each object account in the application platform, and dividing the account interaction network into multiple account communities, the potential influence coefficient between accounts is calculated within each account community according to the push diffusion mode and the interactive operations between each node. Combining the node influence coefficient of the seed account itself and the potential influence coefficient between accounts, the influence coefficient of the candidate account on the seed account is calculated. Finally, based on the influence coefficient, the target account for receiving the target advertising resources is determined from the candidate accounts. In other words, this embodiment of the application accurately calculates the potential influence coefficient between nodes by combining the interactive operations between object accounts, thereby determining the target account that is significantly influenced by the interactive behavior of the seed account, improving the accuracy of the determined target account for receiving the target advertising resources, and increasing the advertising conversion rate.

[0215] It should be noted that the embodiments of the target account determination device here can refer to the embodiments of the target account determination method described above, and will not be repeated here.

[0216] According to another aspect of the embodiments of this application, an electronic device for implementing the above-described method for determining the target account is also provided. This electronic device may be... Figure 1 The target terminal or server is shown. This embodiment uses the electronic device as an example to illustrate the concept. Figure 11 As shown, the electronic device includes a memory 1102 and a processor 1104. The memory 1102 stores a computer program, and the processor 1104 is configured to execute the steps of any of the above method embodiments via the computer program.

[0217] Optionally, in this embodiment, the aforementioned electronic device may be located in at least one of a plurality of network devices in a computer network.

[0218] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0219] S1. Construct an account interaction network using the interaction behavior data generated by each object account in the application platform. Each node in the account interaction network represents an object account, and each edge in the account interaction network represents an interaction operation between two connected object accounts.

[0220] S2, the account interaction network is divided into multiple account communities, where the affinity between object accounts within the same account community is greater than the target threshold, and each edge is configured with a first potential influence coefficient determined based on the interaction operations between nodes within the account community.

[0221] S3, identify seed accounts and push diffusion patterns from the account interaction network. Seed accounts are accounts that meet the push conditions that are compatible with the target advertising resources to be pushed. Push diffusion patterns refer to the way to determine the diffusion of target advertising resources based on the potential influence of seed accounts on other accounts in the account interaction network.

[0222] S4, starting from the seed account, candidate accounts are determined from the account interaction network according to the push diffusion mode. The first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to the second potential influence coefficient.

[0223] S5 uses the second potential influence coefficient and the node influence coefficient that matches the seed account to identify the target account from the candidate accounts to receive the target advertising resources.

[0224] Alternatively, as those skilled in the art will understand, Figure 11 The structure shown is for illustrative purposes only. Figure 11 This does not limit the structure of the aforementioned electronic devices or electronic equipment. For example, electronic devices or electronic equipment may also include components that are more... Figure 11 The more or fewer components shown (such as network interfaces, etc.), or having the same Figure 11 The different configurations shown.

[0225] The memory 1102 can be used to store software programs and modules, such as the program instructions / modules corresponding to the target account determination method and apparatus in this embodiment. The processor 1104 executes various functional applications and data processing by running the software programs and modules stored in the memory 1102, thereby realizing the aforementioned target account determination method. The memory 1102 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 1102 may further include memory remotely located relative to the processor 1104, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. Specifically, the memory 1102 may be used, but is not limited to, to store account interaction networks, first potential influence coefficients, and second potential influence coefficients, etc. As an example, such as Figure 11 As shown, the memory 1102 may include, but is not limited to, the first processing unit 1002, the partitioning unit 1004, the second processing unit 1006, the third processing unit 1008, and the fourth processing unit 1010 in the target account determination device. Furthermore, it may include, but is not limited to, other module units in the target account determination device, which will not be elaborated upon in this example.

[0226] Optionally, the transmission device 1106 described above is used to receive or send data via a network. Specific examples of the network described above may include wired networks and wireless networks. In one example, the transmission device 1106 includes a Network Interface Controller (NIC), which can be connected to other network devices and a router via a network cable to communicate with the Internet or a local area network. In another example, the transmission device 1106 is a radio frequency (RF) module, used for wireless communication with the Internet.

[0227] In addition, the aforementioned electronic device also includes: a display 1108 for displaying the aforementioned scene image and target object list; and a connection bus 1110 for connecting the various module components in the aforementioned electronic device.

[0228] In other embodiments, the target terminal or server can be a node in a distributed system, which can be a blockchain system. This blockchain system is formed by connecting multiple nodes through network communication. The nodes can form a point-to-point network, and any type of computing device, such as a server or target terminal, can become a node in the blockchain system by joining this point-to-point network.

[0229] According to another aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the target account determination method provided in various optional implementations of the above-described server verification processing, wherein the computer program is configured to execute the steps in any of the above-described method embodiments at runtime.

[0230] Optionally, in this embodiment, the computer-readable storage medium described above may be configured to store a computer program for performing the following steps:

[0231] S1. Construct an account interaction network using the interaction behavior data generated by each object account in the application platform. Each node in the account interaction network represents an object account, and each edge in the account interaction network represents an interaction operation between two connected object accounts.

[0232] S2, the account interaction network is divided into multiple account communities, where the affinity between object accounts within the same account community is greater than the target threshold, and each edge is configured with a first potential influence coefficient determined based on the interaction operations between nodes within the account community.

[0233] S3, identify seed accounts and push diffusion patterns from the account interaction network. Seed accounts are accounts that meet the push conditions that are compatible with the target advertising resources to be pushed. Push diffusion patterns refer to the way to determine the diffusion of target advertising resources based on the potential influence of seed accounts on other accounts in the account interaction network.

[0234] S4, starting with the seed account, candidate accounts are determined from the account interaction network according to the push diffusion mode. The first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to the second potential influence coefficient.

[0235] S5 uses the second potential influence coefficient and the node influence coefficient that matches the seed account to identify the target account from the candidate accounts to receive the target advertising resources.

[0236] Optionally, in embodiments of this application, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0237] Optionally, in this embodiment, those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the target terminal. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0238] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0239] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods of the various embodiments of this application.

[0240] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0241] In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0242] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0243] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0244] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for determining a target account, characterized in that, include: By utilizing the interactive behavior data generated by each object account in the application platform, an account interaction network is constructed. Each node in the account interaction network represents an object account, and each edge in the account interaction network represents an interactive operation between two connected object accounts. The account interaction network is divided into multiple account communities, wherein the affinity between object accounts contained in the same account community is greater than a target threshold, and each edge is configured with a first potential influence coefficient determined based on the interaction operations between the nodes in the account community. Seed accounts and push diffusion patterns are determined from the account interaction network. The seed account is an account that meets the push conditions that are compatible with the target advertising resource to be pushed. The push diffusion pattern indicates the way to spread the target advertising resource based on the potential influence of the seed account on other accounts in the account interaction network. Starting from the seed account, candidate accounts are determined from the account interaction network according to the push diffusion mode, wherein the first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to the second potential influence coefficient. Using the second potential influence coefficient and the node influence coefficient that matches the seed account, the target account for receiving the target advertising resource is determined from the candidate accounts.

2. The method according to claim 1, characterized in that, The step of determining the seed account and push dissemination pattern from the account interaction network includes: Based on the account attributes of each account in the account interaction network, an initial seed account set is determined, wherein the account attributes include whether each account meets the push conditions that are compatible with the target advertising resource to be pushed; The seed accounts that meet the preset conditions are determined from the initial set of seed accounts; The push diffusion mode is determined based on the diffusion ratio, wherein the diffusion ratio is the ratio between the number of target accounts determined from the account interaction network and the number of seed accounts.

3. The method according to claim 2, characterized in that, The step of determining the seed accounts that meet the preset conditions from the initial seed account set includes: The evaluation factors for each account are determined by utilizing the social capabilities and account activity of each account in the initial seed account set. The evaluation factors of each account are sorted in descending order, and the top N accounts are determined as the seed accounts, where N is a positive integer greater than or equal to 1.

4. The method according to claim 1, characterized in that, The process of identifying candidate accounts from the account interaction network, starting with the seed account and following the push diffusion pattern, includes: Based on the first potential influence coefficient determined by the interaction between the nodes in the account community and the interaction between the nodes in the account community, the predicted potential influence coefficient between the nodes in the account community is determined, wherein the first potential influence coefficient represents the potential influence coefficient determined when there is an interaction between the nodes; The candidate accounts are identified from the multiple account communities using the predicted potential influence coefficient.

5. The method according to claim 4, characterized in that, The determination of the predicted potential influence coefficient of the edges configured between the nodes in the account community, based on the first potential influence coefficient determined by the interaction operations between the nodes in the account community and the interaction operations between the nodes in the account device, includes: When the diffusion ratio is greater than a preset threshold, singular value decomposition is performed on the target adjacency matrix to obtain a first implicit feature vector and a second implicit feature vector. The target adjacency matrix is ​​a matrix determined based on the first potential influence coefficient determined by the interaction operation between the nodes in the account community. The dimension of the target adjacency matrix is ​​M×M, where M is the number of all object accounts in the account interaction network and M is a positive integer greater than or equal to 2. Based on the first implicit feature vector and the second implicit feature vector, the similarity between the nodes within the account community is determined; The similarity between the nodes is determined as the predicted potential influence coefficient configured for the edges between the nodes.

6. The method according to claim 4, characterized in that, The determination of the predicted potential influence coefficient among the nodes in the account community, based on the first potential influence coefficient determined by the interaction operations between the nodes within the account community and the interaction operations between the nodes within the account community, includes: When the diffusion ratio is less than or equal to a preset threshold, the predicted potential influence coefficient of the edge configured between the nodes is determined according to the influence weight corresponding to the interaction operation between the nodes in the account community.

7. The method according to claim 6, characterized in that, The step of determining the predicted potential influence coefficient between nodes based on the influence weights corresponding to the interaction operations between the nodes within the account community includes: In the event that there is a first interactive operation between the first object account indicated by the first node and the second object account indicated by the second node within the account community, a first potential influence coefficient corresponding to the first interactive operation is determined. In the case where there is a second interactive operation between the second object account and the third object account indicated by the third node in the account community, a first potential influence coefficient corresponding to the second interactive operation is determined. In the absence of any interaction between the first object account and the third object account, the product between the first potential influence coefficient corresponding to the first interaction and the first potential influence coefficient corresponding to the second interaction is determined, and the product is determined as the predicted potential influence coefficient corresponding to the edge formed by connecting the first object account and the third object account.

8. The method according to claim 1, characterized in that, The step of determining the target account for receiving the target advertising resource from the candidate accounts by utilizing the second potential influence coefficient and the node influence coefficient matching the seed account includes: When the number of seed accounts is F and the number of candidate accounts is K, the target influence coefficient of each candidate account among the K candidate accounts is determined by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account, where F and K are positive integers greater than or equal to 2. The target influence coefficient of each candidate account among the K candidate accounts is sorted, and the Q candidate accounts corresponding to the top Q target influence coefficients are determined as the target accounts for receiving the target advertising resources, where Q is a positive integer greater than or equal to 1 and less than or equal to K.

9. The method according to claim 8, characterized in that, The step of determining the target influence coefficient of each candidate account among the K candidate accounts being influenced by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account includes: The influence coefficient of the i-th target affected by the F seed accounts on the i-th candidate account is determined as follows, where i is a positive integer greater than or equal to 1 and less than or equal to K: Determine the influence coefficient of the i-th candidate account on each of the F seed accounts, and obtain F influence coefficients; The largest of the F affected coefficients is determined as the target affected coefficient of the i-th candidate account being affected by the F seed accounts.

10. The method according to claim 9, characterized in that, Determining the influence coefficient of the i-th candidate account by each of the F seed accounts includes: The following steps are used to determine the j-th influence coefficient of the i-th candidate account being affected by the j-th seed account, where j is a positive integer greater than or equal to 1 and less than or equal to F: The j-th second potential influence coefficient is determined as the edge configured between the i-th candidate account and the j-th seed account; The j-th node influence coefficient and the j-th second potential influence coefficient that match the j-th seed account are multiplied together to obtain the j-th influenced coefficient.

11. The method according to claim 8, characterized in that, The step of determining the target influence coefficient of each candidate account among the K candidate accounts being influenced by the F seed accounts based on the second potential influence coefficient and the node influence coefficient matching the seed account further includes: The influence coefficient of the i-th target affected by the F seed accounts on the i-th candidate account is determined as follows, where i is a positive integer greater than or equal to 1 and less than or equal to K: Determine the influence coefficient of the i-th candidate account on each of the F seed accounts, and obtain F influence coefficients; The F affected coefficients are weighted and summed to obtain the target affected coefficient of the i-th candidate account being affected by the F seed accounts.

12. The method according to any one of claims 1 to 11, characterized in that, Before determining the target account for receiving the target advertising resource from the candidate accounts using the second potential influence coefficient and the node influence coefficient matching the seed account, the method further includes: A group of accounts that do not meet the conditions are selected from the candidate accounts, wherein the conditions include that the accounts in the group are the seed accounts and that the accounts in the group are abnormal accounts.

13. The method according to any one of claims 1 to 11, characterized in that, Each edge is configured with a first potential influence coefficient determined based on the interaction operations between the nodes within the account community, including: In the case where there is a first type of interaction and a second type of interaction between the first node and the second node in the account community, the product of the interaction frequency of the first type of interaction and the first interaction weight of the first type of interaction is determined as the first product. The second product is determined by multiplying the frequency of the second type of interactive operation by the weight of the second type of interactive operation. Summing the first product and the second product yields the first potential influence coefficient configured for the edge between the first node and the second node.

14. A device for determining a target account, characterized in that, include: The first processing unit is used to construct an account interaction network by utilizing the interaction behavior data generated by each object account in the application platform. Each node in the account interaction network is used to represent an object account, and each edge in the account interaction network is used to represent an interaction operation between two connected object accounts. A partitioning unit is used to partition the account interaction network to obtain multiple account communities, wherein the association closeness between object accounts contained in the same account community is greater than a target threshold, and each edge is configured with a first potential influence coefficient determined based on the interaction operation between the nodes in the account community. The second processing unit is used to determine seed accounts and push diffusion modes from the account interaction network, wherein the seed account is an account that meets the push conditions that are compatible with the target advertising resource to be pushed, and the push diffusion mode represents the way of spreading the target advertising resource based on the potential influence of the seed account on other accounts in the account interaction network. The third processing unit is used to determine candidate accounts from the account interaction network starting from the seed account and according to the push diffusion mode, wherein the first potential influence coefficient configured on the edge between the seed account and the candidate account has been updated to the second potential influence coefficient. The fourth processing unit is used to determine the target account for receiving the target advertising resource from the candidate accounts by using the second potential influence coefficient and the node influence coefficient that matches the seed account.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program can be executed by a terminal device or computer at runtime as described in any one of claims 1 to 13.

16. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1 to 13.

17. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method described in any one of claims 1 to 13 through the computer program.

Images